CRMtech Archives - Marketing Technology

January 13, 2018July 15, 2018

8 MACHINE LEARNING for marketing areas to watch in 2018

If you’re like the unbreakable Kimmy Schmidt and got stuck in a bomb shelter in 2017, it may be both a blessing and a curse that you missed the machine learning for marketing media frenzy. Machine learning showed up everywhere, rivaling electricity’s systemic emergence a century ago, allegedly injecting sage-like wisdom into everything from sales forecasting tools to email subject lines generators.

But buildup and hype aside, real progress was made in using machine learning for marketing purposes, infiltrating impactful areas as unprecedented investments poured in. More resources supporting great minds pushed forward innovation in areas like image recognition, voice technologies, and natural language generation (NLG). And savvy brands that mindfully wired these into marketing applications boosted performance, in some cases realizing 400 percent ROI. Here are eight areas worth watching in 2018 that saw significant advancement and are well-poised to advance further.

Big data and a need for speed

Like real estate’s mantra of location, location, location…. machine learning’s very foundation and success are predicated on its thirst for big data and its need for scaled-out, lighting-fast processing speed.

But for data lovers, just as the internet giveth, during its unabashed wild-west data rush era, privacy laws spurred on by libertarian outcries soon may taketh it back. So, keep an eye on data privacy regulations, such as GDPR (which takes effect in the European Union in May 2018), as they could seriously impact future data availability.

Regulatory environments notwithstanding, with abundant data stockpiles and processing speeds continuing an inexorable march forward (vis-a-vis faster GPUs and cloud computing), expect more advances. For example, firms will latch onto progressive profiling and incremental data hygiene methods to refine first-party data, as emphasis shifts away from second and third-party data sources subject to stricter privacy regulations.

Capital One did just this in a routine email sent in late 2017, when they requested members update annual income data on file (previously obtained by appending from a third-party source), suggesting that if customers cooperated, higher lines of credit would be their reward.

2018 will see more of this. Organizations will harvest their big data crops, sifting off customer behavior insights aimed at deepening relationships and selling more products faster using less resources. Anticipate more investment in customer data platform, compiling, virtualization, and rationalization initiatives, with more computing horsepower and human capital helping the harvesting efforts in 2018.

Marketers! You need bionic ears & AI voices

As humans, we’re obsessed with creating and perfecting tools that overcome our limitations, take our skills to new levels, and make our lives better. And last year marked the point that AI devices such as natural language processing, text analytics, and language generators stormed the commercial scene and provided marketers with enhanced listening and speaking abilities.

Listening means understanding not just hearing. Enterprise marketing experts were graced with technology that can listen and understand millions of customer inputs simultaneously across a plethora of channels. Call scripts, reviews, complaints, social posts, and a host of other forms of feedback can be ingested, concept labeled, checked for sentiment, and gleaned for intent. Look for more applications and advances that propel the viability of using tech to listen and understand the voice of the customer at scale.

Source: http://www.scmp.com/news/hong-kong/economy/article/2080503/hsbc-launch-voice-recognition-hong-kong-phone-banking

Although Siri, Alexa, Amelia, Cortana, and other AI assistants weren’t born in 2017, they arguably came of age, infiltrating our homes, and entering the workplace. If you didn’t catch it, Amazon announced Alexa for business at its re:Invent conference in November. Machine voices will continue to spread to business places like conference rooms, service channels, products, and kiosks. And companies (such as HSBC, Citi, and Barclays) found voice signatures another reliable biometric authentication tool to streamline digital transactions.

In 2018 machine learning may not replace you, but using it to handle routine tasks, listen to and converse with customers, and accept it as part of your marketing, service, and sales team will be essential to your survival, as you’re asked to up your productivity and customer experience enhancing game.

Put machine learning’s eyes on customer data, journeys, and marketing content

Discovering, understanding, and learning from customer journeys requires mechanisms to observe and quickly answer question such as:

Which customers are eligible for offers, got them, and responded
Where do customers struggle, pause, or get stuck in their journeys
What sequence of offers and channels lead to conversion (attribution)
When do certain customers show up on the marketing radar; and when do some drop off and why

Marketing specialists started using journey analytics to piece together the customer behavior puzzle, and the tech got better at going beyond business intelligence guesswork to prescriptive AI. More AI vendors bubbled up offering solutions that don’t just sum and sort data, but provide an analysis layer peppered with NLG narratives (such as Narrative Sciences and Arria). Others majored on providing better path-to-purchase journey visualizations, like Clickfox and Pointillist (although its arguable whether these are really AI tools).

And some focused efforts at bringing image recognition to real machine learning for marketing use. Deep learning and image recognition applications went far beyond surfacing that labradoodles and fried chicken appear related. AI image processing proved its mettle for filtering and categorizing marketing and sales content, helping marketers better understand customers’ content needs and serve them appropriate and relevant content and offers. Brands began expediting and personalizing services using the ubiquitous smartphone and AI’s ability to pinpoint products and people in pictures and video. For instance, Aurasma launched an app that democratizes adding augmented reality to a consumer experience by simply triggering a video or animation overlaid on a smartphone screen based on recognizing a pre-defined image.

“Hey AI! Create me some emotionally compelling content”

Marketing pros earn their pay by crafting compelling content using words and visuals to express value and elicit responses. They dance their evocative content lures in front of consumers knowing those customers will strike if needs are met and emotions satisfied. But up until just recently, most of these assets were home spun.

Yet last year, avant-garde marketers began applying AI to content generation, realizing that to compete in the new world (where content must be both mass produced and highly personalized), old tools must give way to new ones.

Source: https://blog.7mileadvisors.com/natural-language-generation-the-game-changer-for-the-21st-century-16b5a7ed3336

And firms like Persado began facilitating the march toward marketing’s creative nirvana, using NLG, emotional science, and machine learning to optimize (down to the preferences of an individual) the attractiveness of marketing offers by altering language, font, color, position, and other creative formatting. Results are not just encouraging, they’re somewhat staggering: click-through-rates (CTR) increased by 195 percent; conversion increased by 147 percent.

In one case using this technology, Amex Rewards generated 393,000 versions of engineered copy for its banner ads aimed at getting a customer to burn down their rewards points. The winner drove an 8.6% conversion rate, thumping the control’s 3.5% rate.

Self-driving marketing – Your AI digital agency

Practitioners continue to debate whether machine learning data prep, analytics, and marketing in general can be fully automated (particularly at the enterprise level), but nonetheless, the tools keep coming.

To this end, an interesting arrival on the scene was a vendor called Frank.ai, albeit clearly for down-market marketers. It’s literally 8 steps to setup and run a multi-channel campaign:

Enter name and dates for campaign
Select audience by city, interests (mix of music, pop culture, shopping, sports, etc..) or look-a-like targeting; age (typical bands); gender; language
Decide on display ad on desktop or mobile or both
Specify budget (e.g., $1000)
Upload display ad creative image
Add social media promotional ad (if desired)
Add URL for click through (analytics tracking automatically setup in Google Analytics)
Enter payment method (credit card or PayPal)

Simple and unsophisticated? Check. Will this kind of tech put further pressure on enterprise vendors to make their tools easier to use? Check.

Explainable machine learning for marketing

As machines crunch data, score customers, make predictions, and automate marketing, being able to explain to humans what’s going on and why is becoming more important.

Some models are very opaque, and simply can’t explain themselves. Given this, firms will need AI controls in place (such as offered by Pega) to prevent opaque models from being deployed in certain situations. Others are more transparent, easier to tease apart, and safer to unleash. Research and applications are stepping up in this area, so stay tuned, especially as more regulations emerge such as GDPR, that dictate data rights and demand algorithmic transparency.

Building one machine learning for marketing brain

Like opinions, everyone seemed to have an machine learning software brain to peddle in 2017 including:

Watson from IBM
Einstein from SFDC
Sensei from Adobe
DaVinci from SAP
Magellan from OpenText
Always-on Customer Brain from Pega

What was less clear, however, was if each had one coherent well-integrated brain – or instead a multitude of disparate intelligence modules from the various acquisitions. In the case of SFDC, for example, between 2012 and 2016 they acquired 21 companies, of which at least nine had some form of machine learning for marketing tech.

Stay tuned to AI developments from these and other leading marketing technology vendors, and pay close attention to whether they demonstrate real intelligence integration in the solutions they sell.

Machine learning for marketing organizational dynamics

Accomplished scientists and artists have rarely been cut from the same cloth. In 2017, Walter Isaacson released his long-awaited masterpiece, the biography of Leonardo da Vinci, adding it to his corpus of history’s best examples of exceptions to this rule (Ben Franklin and Albert Einstein being other similar biographies he’s written).

So rather than wait for enough Leonardo types to come along, organizations would be wise to work toward making connections across machine learning and creative disciplines, which will be key to maximizing their capacity to innovate.

Along with attracting, merging, and retaining the right talent, brands must also acquire the right machine learning technology, but even more important is having a concerted AI strategy closely coupled with business objectives and marketing improvement goals. It’s imperative to work from well-defined use cases and clearly articulated outcome definitions backward to the technological and data solutions necessary to support them. Further, firms must use nimble organizational structures with small teams made up of artists and scientists; IT and the business; re-aligning resources into small digital factory teams that are wed to agile methodologies and collaborative approaches.

2018 and beyond

In all, 2017 was a banner year for machine learning for marketing, in terms of both hype and legitimate commercial progress. Keep track of these eight areas, and you’ll be following the most interesting and promising leading-edge AI technologies and trends that will prove paramount to success in improving and automating marketing and customer experience.

November 27, 2017July 15, 2018

Customers Are INDIVIDUALS Not Averages | How RTIM Treats Them Special

Real-Time Interaction Management (RTIM) delivers personalized experiences to people.

Earlier this year, I signed up for a points program with a large hotel chain, and somehow my last and first name were reversed in the enrollment process. The next day I noticed the welcome email message started with, “Jeffs, we’re so happy you joined the fam.” Figuring it was my botch I went online and fixed it in my profile – problem solved – or so I thought.

Apparently the erroneous data instantly had spread, like a venomous bite, and propagated to other databases. My feeble attempt to suck it from the source was too late and didn’t work. Still, nearly a year later, I still get messages starting with, “Hello, Jeffs” which rather than setting an intimate tone for the oncoming interaction, sets a grating one. I may still read on, yet I’ve been reminded upfront I’m essentially a bunch of bytes to the interactor on the other end.

I get it – mistakes happen; systems are stitched together, and people (and the systems they use) are under enormous pressures to share data, scale, and automate. Nonetheless, when firms chose to operate this way – neglecting to fix little things, they’re failing to measure the impact of the most fundamental flaws that often make or break an entire customer relationship.

Traditional Marketing Technology (Martech) vendors espouse solutions allegedly providing personalized communications. And when their clients deploy these systems, they assume they’ll develop meaningful relationships with individual customers but, the fact is, most won’t. Customers still routinely report broken processes (like my example), one-size fits all treatments, non-individualized experiences, and very few (only 27 percent) think AI will help.[i] So, if you’ve been entrusted with helping achieve loyalty-building relationships, that’s more than a little discouraging, since it’s not for lack of will, good intentions, invested time, or resources.

As consumers, we browse, research, shop, and purchase constantly – sharing our information freely (sometimes unbeknownst to us). We surrender our identity, intentions, preferences, history, location, and so on – often repeatedly, yet we see little in return in terms of well-tailored products, services, and experiences. And this goes beyond the obvious, such as in self-service experiences (where our expectations for personalization are low), into human-assisted channels where our expectations are higher, but paradoxically we often encounter robotic-like agents.

Consider how brands place customers into huge buckets that dictate treatments:

Most loyalty programs have about four tiers. If a program has 10 million members, that’s about 2.5 million members per tier.

When a data scientist builds a decile-based RFM model (RFM stands for an algorithm that scores based on recent transactions, frequency of them, and their monetary values), that’s 10 segments, and again about 1 million customers per segment.

And even when zip code level data is used, that’s still about 8,000 people to a segment – and let’s face it, as much as you love your neighbors, you know how different you are from them.

Rarely do we enjoy being stereotyped. When we’re assigned to a troupe, and approve of it, it’s usually because we made a conscious choice. We find more differences than similarities when we are forced into artificial groupings, and we get rightfully grumpy with being pigeonholed. Conversely, we rave when companies celebrate our uniqueness, and we love to tell these stories.

RTIM – Your AI ROI Machine

AI, arguably the most overused and abused word of the year, particularly among Martech vendors, does have in its midst the underlying technology to begin to solve for improving and individualizing customer experiences, and in techno-geek terms it’s known as RTIM (Real-Time Interaction Management). Businesspeople using RTIM, however, would rather focus on results versus names, and the reality that these systems consistently generate 300 percent plus ROI [ii] – in other words, they are AI ROI Machines.

Why do RTIM systems outperform traditional marketing automation systems? Simply put, it’s because they make decisions one individual at a time, hence delivering one-to-one interactions. Figure 1 depicts the difference between many of today’s Martech systems and an RTIM system:

Figure 1 – Typical Martech system versus RTIM system

Figure 1’s top lane depicts how traditional Martech systems place customers into segments, assign offers to those segments, and execute treatments in each channel. On the other hand, RTIM systems act on behalf of each customer (instead of tranches of them globed into segments). Moreover, RTIM systems operate based on one set of coordinated rules and analytics linked into a set of arbitration strategies – for thousands of customers per second.

For example, a company with five products marketed by five different divisions uses a single decision engine to resolve the best thing to do for the customer. Consequently, an RTIM approach enables a brand to act as one organization instead of many disparate companies with dozens of conflicting rule engines.

RTIM-based systems recall an individual’s history – instantaneously – each time a customer interacts, factor in new (contextual) information, and calculate the best action. They execute real-time analytics to determine an individual’s propensity to respond to a candidate list of eligible offers, then consider customer value and the economic benefits of the offers before rendering a final decision. Granted, they don’t have perfect knowledge of the person, yet just like a human brain, they remember past interactions and learn from them, and place a premium on the most up-to-date information. They’re also agile enough to perform dynamic recalculations (in less than one second) to further improve the pending decision and enhance the relationship. Through this two-way iterative approach, they essentially carry on a real-time conversation in a single session as shown in Figure 2.

Figure 2: RTIM’s iterative two-way conversational approach

Regardless of the superficial popularity and obfuscation of the term AI, it’s incumbent on us as marketing professionals to inspect the value-added by the underlying CX technology. Earlier this year, Forbes did just that, citing the Forrester Tech Radar on AI technologies, which found decision management as the top hot AI technology (Figure 3).[iii] And decision management is the central capability embedded in RTIM systems.

With RTIM and its decision management, brands can personalize in real-time, improving on legacy and static Martech systems and processes, and reinvent how customers are treated. Decision management enables companies to make analytically arbitrated evaluations during every customer contact, treating each person based on their constantly changing context, fluid needs, and demands for relevance and continuity.

In his article, What is RTIM, Barry Levine calls it “Right Now Contextual Marketing” and goes on to cite work by Forrester analysts Rob Bronson and Rusty Warner – who both helped establish RTIM market awareness that culminated in the RTIM Wave [iv]. Levine describes how RTIM enables marketers to perform “a continual negotiation — a kind of dance — happening in real-time between all available data and all available offers/actions.”

Figure 3: Forrester AI Tech Radar

Sounds obvious, and aren’t brands already doing this? Well, not really. Consider that on many channels:

You see the exact same style screen and get messages identical to those of every other visitor.

You can’t set and store preferences or alerts for receiving communications.

Call center, branch, and store agents seem ill-equipped to set, store, and recall even the most basic details about you, like how many children or pets you have, and what their names are.

When you receive an email, it’s maybe one of a handful of different versions, so again, if the brand sending it has a list with millions of email addresses, you’re receiving the same content as thousands of others.

Advertisements stalk us for products we just bought or already own.

When you place a call into a service center for the 10^th time in 2 weeks (and you’re feeling obligated to invite them to dinner because how much time you’re spending with them), it’s clear the vibe from the agent isn’t exactly a personal one.

When you start a process on one channel and bail out, and then later reconnect, you’re forced to repeat steps.

As marketing practitioners, we can do better. And as with any road to improvement, it must start with an admission that issues exist and they are negatively impacting others.

Impactful Customer Engagement

Great customer engagement starts with customer understanding. And tiny details matter – things that on the surface seem trivial, although later may turn into a customer testimonial like this:

“Yes, that pet store remembered me, thanked me for my business, remembered my dog’s name breed, and age (Sandy, our Westie, is 14 now). They seemed genuinely concerned for her health and status. They provided me valuable insights into her dietary considerations, and their app sends me reminders for refills that I might otherwise miss.”

Unpack that and contemplate what it’s implying – a memorable personal experience with an aura of care, empathy, and value. Remembering names, age, past purchase history, applicable products – admittedly is basic stuff. But take that basic data, and in combination with other factors, use it to systematically treat each customer’s situation uniquely, and you’ll put information and technology to beneficial use.

So, forget whether this approach is using AI or not, and start worrying about whether what it’s doing makes common sense for your customers at the moment of interaction. What matters is not whether you store these details, but whether the underlying technology mines that data, learns customer preferences – and with each transaction gets smarter about optimal timing and consumption patterns, and realizes when to trigger meaningful messages.

Frontline staff are already busy, and they’re increasingly asked to be super-human and to provide white-glove treatment at scale. To do it, they’ll need support from technology that stores and surfaces critical insights at the right time, so they can buck the tendency to treat customers as averages – because, what you don’t want are segment-oriented attitudes like this:

Hey since winters coming, everyone needs a coat so we’re pushing winter parkas.
She’s one of my older fixed income retiree types – they all love that annuity product.
Millennials love iPhones, and tweens always buy that pink Otter case.

You want individual-oriented sentiments:

That was Jim and he’s 62, and you’d never believe that Jim loves ziplining, has a Shih Tzu, and listens to Dubstep late at night while he reads his email.

Rosemary says she’ll never retire. She loves her job, loves to day trade, reads email at lunch, and will likely work for her entire life.

Yes, Sara is a unique. She’s 21 and never responds to text messages, unless from close friends; she gets up early, reads email before work, and is into Hello Kitty, guinea pigs, and Thrash Metal.

Each one of your customers are unique people, not customer id’s in cluster codes. Treat them as such.

Continuous CX Improvement

Back to our little story of the inverted last name. You’re probably wondering, couldn’t that company have solved this problem without an RTIM system? Maybe. But outfits that work from a common customer database, understand the true meaning of “system of record” and synchronize data when its distributed, and use RTIM to operate from an organized set of rules and analytics and make the best possible decisions in the moment are much more likely to consistently get CX right, and to improve it – one person at a time.

[i] What Consumers Really Think About AI: A Global Study, https://www.pega.com/ai-survey, 2017

[ii] Forrester Total Economic Impact (TEI), https://www1.pega.com/insights/resources/forrester-total-economic-impact-tei-pega-marketing, 2016

[iii] Forbes, https://www.forbes.com/sites/gilpress/2017/01/23/top-10-hot-artificial-intelligence-ai-technologies/ , 2017

[iv] The Forrester Wave: Real-Time Interaction Management, https://www.forrester.com/report/The+Forrester+Wave+RealTime+Interaction+Management+Q2+2017/-/E-RES136189, Q2 2017,

August 19, 2017

What it Takes to be a CX Transformer

Thirty years ago, when I unpacked my first computer, a Commodore 64, rigged it to my 13-inch tube TV, and wrote my first program, the process of creating a digital experience hooked me. That I could design and assemble mere bits and bytes, package them up into an asset, refine it, and eventually share it for the benefit of others – for entertainment or problem solving – just enthralled me.

With time and market efficiency sorting who gets paid to do what, I altered my path away from programming and toward design and consulting, leaving the coding and compiling jobs to those more talented than me in that trade. That wonderful feeling of accomplishment, however, never left me and still drives me today. Whether it’s creating visual concepts, designing software, or producing media, creating a re-usable asset with experiential worth (striving to be a CX transformer), for me, is a universal and time-tested motivator.

Experiential assets, originally made from scratch, must evolve to the liking of their benefactors. They invariably play a role in nearly every commercial experience. For example, a vehicle manufacturer produces a physical product, but the agency who markets it as well as the dealer who sells and services it – all add crucial elements into the customer’s journey of shopping for, buying, and owning that vehicle – all contributing to (or subtracting from) accumulated impressions of overall worth and value.

Organizations are either born with this mentality, where it’s baked into the fabric at every level and function of the organization, or they must transform. Startups who don’t adopt this mentality burn through money and soon dissolve. Legacy firms are faced with odds not unlike that of a recovering addict. Most hit bottom, before they realize the extent of their problem, and by then it’s often too late. Few are afforded the chance to recover and most who try will regress. In fact, a recent Forrester study [i]indicated as many as 77% of those who embark on CX transformation will fall short.

With all this buzz, don’t we already get great CX?

The short answer is, not really. According to a global survey [ii]of 7000 consumers, 89% “think brands need to work harder to create a seamless experience for customers.” There’s lots of talking about seamless and personalized experiences, and less walking the walk. And consumers continue to report a deficit of it, as evidenced in an Infosys survey [iii]indicating that 73% have never experienced online personalization. Here’s the reason: Many of us, and the firms we work for, aren’t practicing what we preach.

Regardless of what you do, you’re in the business of creating customer experiences. Whether in sales, marketing, service, or operations; whether you set vision, do design work, code, implement, consult, or the like, your ultimate mission is creating something that someone else appreciates and finds value in – because it makes their life better. It improves their experience. If you can’t tie what you do and why you do it back to that, your mission is misdirected.

The only reason customers buy, use, or recommend products or services is because they experience value. So, if you simply blabber about CX but don’t improve it, you’re subtracting value, like in figure 1:

Figure 1: All Talk Equals Value-Subtracted from CX

Everyone plays a role in experience management. For example:

If you’re a banker, during any interaction, clients are judging each aspect of your services. When they point out friction, dissatisfaction, annoyances, frustrations – they aren’t being pests – they’re handing you gold.
If you manage a telco’s call center, though one step removed from direct feedback, front-line agents will hand you that gold. Will you ignore it, or will you investigate, catalog it, document it, and act on it?
If you design software used by that banker or agent, you’re instrumental to how the total experience comes off when moments of customer truth occur. Software augments customer facing CX delivery, either enhancing it or contributing to its malfunctions.

Software and AI technologies have already changed our lives, and continue to transform how we experience life. From when we wake to the minute we doze off, the way we interact with the world, for business and pleasure, is vastly different now from the day I cracked open that Commodore box.

Data is abundant and the right intelligence in software is available. Yet how both are captured and deployed is what spells the difference between memorable moments versus forgettable incidents. Dated advice, cloaked as sage recommendations, abounds on what data to tap and which AI technologies to trust.

Beware of the CRM “Catchers in the Rye” who have a vested interest in selling old software disguised as AI and one-to-one personalization, spruced up with fancy new names like Customer Data Platforms, but stuck in a forgone era. Peel these back and see if they rest on an old batch and blast architectures with no real proven use cases for predictive analytics, built essentially for pushing emails to segments. You’re sure to hit a wall with these, since they were never built to handle real-time, analytics based one-to-one contextual engagements. If you’re interested, I cover this topic in more depth in this article.

Or worse still, beware the do-it-yourself CRM & AI pushers, selling piles of new programming gadgets with exotic names such as Python, Storm, Spark, and Kafka, but missing the warning label that says, “Much assembly required.”

CX Transformation Process

The transformation process, contrary to overhyped tales of sudden disruption, is mostly evolutionary. It involves creative minds with an unwavering and relentless obsession to improve experiences – as measured by customers. But today you must do everything you can to go through this process fast.

Iteration (figuring out how to improve) means executing various steps in succession – speedily and repeatedly to learn fast. It also takes a flexible methodology and tools supporting rapid revisions. Each time Thomas Edison’s filament didn’t work, he wasn’t failing, he was learning. When asked about racking up so many failures, Edison replied, “I have not failed 10,000 times. I have not failed once. I have succeeded in proving that those 10,000 ways will not work. When I have eliminated the ways that will not work, I will find the way that will work.”

Be unyielding in finding gaps, filling needs, overcoming shortcoming, and plugging them with an improved asset. Find the simple stuff, that exacerbates customers, but is easily addressed. Do ten thousand little things right – and fast.

To succeed, you’ll need to be well-equipped with the right CX transformation methodology and technology. Speed to market and economies of scale matter now more than ever. It takes steadfast customer centric vision, modern tooling, and an agile methodology. Let’s explore the four key steps shown in figure 2.

Figure 2: Depicting the CX transformational process steps

CX Transformer Step #1: Conceive Innovation

As you come up with a concept, consider the objectives…. making things better, faster, cheaper. Ideally, you’ll eventually address all these, but practically you’ll need focus. Will the proposed innovation fix something that is terribly broken? Better yet, will it preemptively address a shortcoming. Often, fixing inadequacies is simple, yet the consequences of not fixing them are huge.

To find opportunities for CX innovations, use analytic heatmaps fed by behavior data on websites and mobile devices to zero in on where customers struggle or bail out. Mine reviews, comments, call logs to find repeating themes.

Here’s an example I heard from a person I sat next to on a flight. He had booked a trip to Dubai, but the travel service never proactively alerted him that travel to UAE requires a passport that doesn’t expire in less than six months. On his departure day, he couldn’t check in, and subsequently was on the phone for hours, working the problem and seeking amends for this horrible experience. The root cause was recorded in logs. The fix (innovation if you will) was rudimentary and excruciatingly easy.

“If customer books trip to country X, and passport expiration date is Y, alert customer about passport rule.”

In this case, the customer placed a gold nugget into the lap of the brand, begging them to fix it for future customers. Will they? Only if they’ve institutionalized collecting hiccups like this, and weaving them into the innovation and improvement process.

Think of innovations in sets. Will the CX innovation set be press-worthy; will the total experience be unique and better? Take the innovation set and break it down into manageable chunks. To improve service usability, for example, consider whether the specific design is elegant, visually appealing, modern, stylistic, easily navigated, intuitive, and so forth. Remember, even when just creating a form, such as an insurance policy application, all the above matters in CX.

Spend three times as much effort on design versus construction. If service improvement is your aim, pick (as your innovation set) a critical customer journey that cuts across various functions and channels, and obsess with its design. While iterating on the design, always apply a range of customer sniff tests tied to customer personas. How would customer X use this? How would customer Y perceive this?

Just as incentive drives employee behavior, it drives customer behavior. Customers are motivated by the value they both perceive and achieve from using your products and services, regardless of the organizational excuses they encounter along the way.

CX Transformer Step #2: Judge Harshly

Critique innovations, not just with self-criticism, but with the varied feedback of others. Compare to market alternatives and what big competitors are doing and what customers complain about. Once again, view the current state of the experience through customer eyes. Clients not only measure success, they also give clues about required innovations. If an asset works they use it, open it, share it, like it, and buy it.

Watch exactly how customers use the innovation. Designers call this usability testing, and too often, it’s shortcut out of the development process in the name of speed. Watch how customers interact, how they shop, how they decide, whom they consult with, and why they buy. Look for where they struggle, the questions they ask, why they need help, and ask what went wrong. Then go back to the drawing board to create a new experience, craft a new email, create a form, redesign a web page, or work on ideas to improve how agents engage with customers.

Use a basic four quadrant Risk / Reward matrix, as shown in figure 3, to prioritize a backlog of CX improvement opportunities.

Figure 3: Risk (Effort) / Reward (Value) matrix used to prioritize innovation ideas

Don’t make your goal mimicking competitors, but instead to gauge your inferiorities to them, study their winning ways, and chart your course – but dare to be different – then test and learn. Compare your asset to others available in market. This guides, both in terms of whether you’re behind, but also what hasn’t been done – thus presenting opportunities to do something new, something unique.

Pattern yourself on proven winners, not just in your industry, but also in very different ones. Why? Because that’s where unique ideas come from – not from copying your competitors, but from proxies that when applied to a different problem become a new idea.

For instance, to transform the branch experience for its customers, Capital One recently introduced café style locations, drawing on a combination of Starbucks and Apple store concepts.

CX Transformer Step #3: Apply a Value Test

Determine whether your innovations improve experience. To do this, perform behavior tests and not just surveys. People don’t always do what they say they’ll do. Test your innovation by getting real customers to use it in production pilots, and then measure whether, for instance, the task was accomplished faster.

Getting there may not be easy, cheap, or fast, but if your product isn’t passing these tests, you haven’t improved your customer’s experience. Each innovation should pass at least one of these tests, and collectively overtime, it must pass all three.

At this stage, the test is if your customers are buying or using your asset. If they see value, they’ll do these things, so measure for it, and use this as your ultimate yardstick.

CX Transformer Step #4: Analyze Objectively

Once you release your concept into the memorialized world of production, objectively (and recurrently) evaluate its worth. What works today may not work tomorrow. In addition to pure customer feedback, consider getting an objective third party to scrutinize it, since creators as well as customers have blind spots and biased views.

For all its advances, and there are many, CX today – when analyzed objectively – is still mostly choppy, dysfunctional, too slow, and places too much burden on the customer. Admittedly, some industries (such as banking and telecommunications) have made more progress than others, yet largely, especially for massive enterprises, CX is frankly still very siloed.

Firms spend millions of dollars on data collection, design thinking, journey mapping, voice of customer, CRM systems, employee training, and so on. Yet when these efforts are not coordinated around a systematic process, data, technology, and culture – hyper coordinated and committed to improving CX –most of that investment will be for naught.

It’s human nature to either ignore feedback or want to defend your baby’s looks, and if you’re busy defending versus fixing simple things, CX won’t improve much. It’s also human nature to pass the buck – meaning no one will take responsibility, because even though at our core we’re pack animals, it’s ironically not in our nature to communicate issues across organizational pillars.

CX transformation doesn’t come easy and it doesn’t come cheap, and rarely comes fast. But for those who listen to and watch customers, fix ten thousand small things fast, live by the adage innovate or die, and cross-functionally collaborate on behalf of better customer experience, the rewards will be plenty.

[i] Forrester, http://www.datastax.com/wp-content/uploads/resources/whitepaper/Forrester-CX-TLP_DataStax.pdf, April 2017

[ii] Zendesk, http://d16cvnquvjw7pr.cloudfront.net/resources/whitepapers/Omnichannel-Customer-Service-Gap.pdf, November 2013

[iii] Infosys, https://www.infosys.com/newsroom/press-releases/Documents/genome-research-report.pdf, 2013

June 7, 2017June 7, 2017

Revolutionize CX with Real-Time Contextual Engagements

When we’re on journeys, we often encounter decisive moments, and how fast we get accurate and relevant information and the decisions we subsequently make can have a huge impact on the outcome. Take, for example, one of your vacations. Your airline cancels a flight leaving you precious seconds to react and find other options before they evaporate; Or, if you’re upset with your wireless service and decide to reach out – calling to give them one last chance. Every case is different in terms of the people involved and the situational context, but timing and the flow of relevant information always plays a crucial role.

Time Is Of the Essence

If you’ve ever managed a project or worked with a project manager, one of the first things you learn is what’s called the triple constraints: the scope (quality), the cost, and the time (schedule) for a project. Project managers always manage a project using this idiom:

“No problem, I’ll deliver as long as I can control at least one of these 3 constraints. Deliver by Jan 1 you say. Fine it will cost X and we can deliver Y scope. Change the scope and the cost will go up. You want to hold costs to $50k, then we’ll deliver X scope by Jan 1. Change the scope and the date will slip.”

It seems lately in projects and commercial transactions we’re trying to break these laws. We want everything faster, cheaper, and more of it at higher quality. Why not, it sounds great. Although the quest for instant gratification isn’t new, it certainly has upped its game.

It’s obvious why we want things faster – because time is a scarce resource and finishing something can mean the difference between opportunity and obsolescence. Since we can’t pack more minutes into a day, we squeeze more tasks into limited time usually sacrificing scope / quality, and all in an attempt to push things to faster completion. With continual pressure to increase productivity levels, technology helps accomplish more with the same resources. Clearly, time is of the essence.

However, there is another reason time is so precious. Time is the enemy of information’s value to decision making. When something happens, information about it may be extremely valuable seconds after it occurs, yet through time that value decays (see Figure 1) at rates that are different depending on its importance to a given decision.

Figure 1:

For example, when someone has a heart attack an enzyme called creatine kinase (CK) appears in the blood. Detecting the recurrence of that enzyme and passing that information quickly to a doctor could save that person’s life. Much later, that information may still be valuable in a post analysis of that patient’s situation, and years later, it could still prove valuable to researchers performing wider studies on congestive heart failure. Nonetheless, its value diminishes considerably over time, eventually relegated to but one data drop in a vast data lake, and ultimately some information even becomes a cost and liability with no corresponding value.

In CRM, real-time data can be of corresponding value and the context of the situation determines its value and decay curve. While on buying journeys, consumers come to decisive moments. When considering what consumers expect at those moments, brands must assume they are impatient, easily distracted, and likely to act suddenly. Once again, speed becomes paramount.

When Do Customer Interactions Need To Be Real-Time?

Honestly, there’s no such thing as real-time information. When an event occurs, time immediately elapses so, at best, a device can detect an event and pass that as information in near real-time to a decision maker. In some cases to be of optimal value, this information needs to reach the decision maker and the decision made – all in less than 1 second:

Given that understanding, what’s most important is that the information makes its way into a decision management system that can factor this into making a decision and in enough time to affect some other event that hasn’t yet occurred.

In our heart attack example, the presence of the enzyme CK is the initial event followed by some device that detects it, followed by that information passing to a decision maker who takes action to prevent a subsequent heart attack:

For CRM, a customer may be on the verge of switching service providers and the event may be one last contact with the existing provider:

Depending on the problem at hand, having real-time information and using it in a dynamic way may be critical to positively affecting customer experience and satisfaction.

In general, customers are less willing to wait for service or answers than ever before. When someone decides to research a potential purchase and browses a website, if it’s not responding nearly instantaneously, it’s likely the consumer will move on. Moreover, if the customer calls into a contact center and enters a long wait queue, again it’s likely the customer will balk. In other words, you literally have seconds to make the right impression and provide them with relevant content.

Situational Context and Relevance

As we’ve just seen, the situations the events occur within, and how events relate to each other are crucial to understanding the full story. The circumstances that form the setting for one or more events are its context, and the data that comes from them contextual data.

Feeding this data into a decision engine can be essential so decision makers understand the full context of each event before making subsequent decisions.

Take another use case. Again, the event is a call coming into customer service. Here’s the setting for that event:

On a rainy day on May 22, 2017 at 4pm, while stuck in a traffic jam on 1-81 in southern VA, a certain customer (Mary) places a call to her credit card provider. Having just learned of late fees, she is not happy. She’s a loyal customer of 23 years who always pays on time and her aim is to get these fees removed.

Consider some of the contextual data in this scenario:

Historical background of customer – e.g., family history, purchase history, engagement history – i.e., knowledge of this customer as an individual (23-year loyal customer – Mary)
Date / Time (May 22, 2017 at 4pm Eastern)
Location / Proximity (northbound on 1-81 at x lat / y long)
Environmental conditions – e.g., weather / traffic (raining – stuck in traffic)
Preferences (uses phone to call in)
Emotional state (unhappy with fees; may be further agitated by weather & traffic conditions)
Journey stage (just received bill and is inquiring about fees)
Current intent / agenda (aims to get fees removed)

Think about how important each piece of information is to the scenario and how all of it, when factored together, forms a more complete picture. For instance, if the call started at 4pm and the conversation with Mary is still in progress at 4:30pm, it’s likely she’s antsy and probably tired from a long day. Now combine that with the weather, traffic conditions, journey stage, current intent, and historical background, and now it’s very likely her patience is wearing thin.

In short, situational context answers these questions about an event.

Situational Context:

Brands that can factor in this kind of contextual information are more likely to address immediate sensitivities and handle customer interactions accordingly. Human agents, unlike artificial ones, have the potential for very high levels of emotional intelligence (EI) [i]but still need assistance compiling the information and making sense of it. So far, artificial intelligent (AI) bots aren’t that advanced in EI and humans have the edge in relating to people, applying soft skills, and making complex judgments. Thus, it’s crucial to realize the full situation when designing systems that will include both self-service options and escalation to humans, and when humans are involved, providing them with augmented information so they can serve customers better.

In this scenario, as with most others, the contextual information becomes practically useless the next day. Unless the systems involved can detect these events and the context in near real-time, and pass it on, the value of this information decays rapidly. Imagine the customer service person saying, “If you just call me back tomorrow morning Mary, my system will run an overnight batch job, and I’ll have a sentiment and value score for you so I can help you better” – said a successful customer service person never.

And systems can’t pre-calculate conditions or recommendations either, hoping that works. How would you rate a pizza delivery company if you ordered a pie, and delivery came in 15 minutes (a near real-time pizza order), but they cooked it the night before?

One Contextual Brain Making Real-Time Decisions

When businesses engage with customers – or vice versa – such as in the examples given, they’re expecting an ongoing conversation regardless of the initial channel. In some cases (such as a phone call), the exchange begins and ends on the same channel, with stimulus and response occurring dynamically and repeatedly. In other cases, such as a website visit or chat session, conversations may take longer, and experience periods of pause, delay, or even happen over multiple channels and sessions.

Unfortunately, many brands treat each interaction as an isolated transaction instead of viewing them as parts of a continuous conversation that builds toward a deeper relationship. When viewed from the latter lens, each interaction nudges the relationship in a positive or negative direction, in effect adding to a “Good Graces” balance, or subtracting from it (see Figure 2). Clearly, the goal of any customer-centric minded organization is to evaluate each interaction’s contribution, monitor this balance, and work to strengthen the relationship – improving it with each engagement.

Figure 2:

This is more than a loyalty point balance, an NPS score, or a sentiment score. It’s more like a credibility score, measuring the overall standing your brand has with a given consumer.

Consumers reward companies with more business when they provide convenience and simplicity, can be trusted, and help improve their lives. Amazon is a prime (pun intended) example. Consider the Amazon Dash button, which really isn’t the revolution in buying experience originally promised, at least not yet. This is the internet-connected button you place, for example, on your washing machine, and you push to reorder detergent. However, many customers purchase the button (for $4.99, which comes with an equal credit upon the first purchase) but never use it. They just love Amazon, trust them (i.e., they have a high good graces balance), and go for their innovative ideas – and this idea smacks of convenience, simplicity, and real-time engagement.

Customers also keep score in terms of brands that just get it. What this means is buyers notice if they get superior, well-coordinated services, regardless of the channel, time, location, or representative involved. For a business to do this, it must have a culture that oozes of a customer-obsessed mentality and have systems that support its people in providing outstanding services at scale. If those systems are in silos (see Figure 3), each commanded by a different brain, achieving that goal becomes impossible.

Figure 3:

Can Anyone (Or Any System) Do This Really Well?

The short answer is yes, companies should be able to do a decent job of this, but they aren’t. Amazingly, a DataStax commissioned Forrester survey of 206 organizations across four major countries revealed, “95% of organizations are currently unable to make sense of customer data, and struggle to gain real-time insights from this data.”

When agents serve customers, to provide great service, they need technological assistance just as a pilot needs instruments. Flight deck instruments must depict actual conditions, with little latency, or they’re useless. Finding out too late that your altitude is 2000 feet could spell disaster. When on chat or a call, consumers expect immediate responses to each question and have a threshold for the total time they’re willing to devote. In theory, whether the agent is a bot or a human makes little difference, what matters is whether the agent provides authentic empathy, good judgement, relevant responses, and resolves the situation quickly.

Today, offloading some of this to machines is possible (sensing conditions, detecting patterns, managing decisions, and triggering actions), but seemingly only with CRM rocket surgeons behind the curtains.

That shouldn’t be the case, and it’s holding back brands from providing exceptional customer experience. Some vendors, nevertheless, are working to change that and provide tooling that allows for real-time interaction management or RTIM[ii] controlled by business users. Though the right tooling alone won’t bring success, as firms must align data, organization, and processes also, choosing the wrong technology will introduce barriers and rigidity that will be difficult or impossible to overcome.

When considering the entirety of the data, insights, and technological solution needed, keep in mind that each vendor has evolved their platform from very different starting points. Some have cobbled together solutions by acquiring smaller firms and in fact don’t even have a single platform. Others have built out their solution, appending capabilities to legacy outbound and batch oriented architectures (often based on email campaign blasting), making it nearly impossible to handle enterprise real-time processing requirements. It’s likely you’ll hit a wall (see Figure 4) with these types of solutions in attempting to transform to real-time 1:1 contextual engagements.

Figure 4:

Thus, look instead for help from partners that have extensive experience with real-time customer interactions and have built solutions from the ground up to handle the complexities of real-time data wrangling, dynamic analytics, and scaling for the instant gratification response times required by today’s customers.

[i] https://hbr.org/2017/02/the-rise-of-ai-makes-emotional-intelligence-more-important

[ii] https://www.forrester.com/report/The+Forrester+Wave+RealTime+Interaction+Management+Q2+2017/-/E-RES136189

April 26, 2017May 18, 2017

10 Commandments of Customer Experience (CX)

This is my shortest post by far. I received these 10 CX thoughts last night in a dream. When they came unto me, they seemed self-explanatory, so I saw no need to elaborate or provide examples. Full disclosure; I did have several adult beverages before I went to sleep:

Strive to know your customers as you would know yourself.

Thou shalt be “Customer-Centric” and put no other products, services or stakeholders before thee.

Thou shalt not make any graven image of customers, such as idol segments. Instead, thou will treat customers as individuals with personalized touch.

Thou shalt not spam customers by carpet bombing with frivolity (causing them to take names in vain).

Thou shalt not contact customers on Sunday…or any day for that matter, unless given permission and there is a relevant service or offer to discuss.

Thou shalt be empathetic and listen to customers, and act with fairness.

Thou shalt not kill off customers with WMDs – “Weapons of Math Destruction” – such as artificial intelligence (AI) algorithms with bias.

Love thy customer, their loyalty, and their journey, and calculate a true LTV (Lifetime Value), not just a year’s worth.

Thou shalt not steal profits from the Customer Innovation Till. A tithe of earnings will be put in said till for pursuing true innovation.

Thou shalt not covet thy customer’s wallet or share of wallet. You will get yours if you obey the other commandments.

April 7, 2017February 3, 2020

Event-based marketing (EBM) & CEP use cases for CRM

Updated: February 3, 2020

Introduction – The Key Event Types

First, check out my latest article on this topic recently published: Shush – Listen for customer signals with event-based marketing & service —- Out of that article surfaced these distinct event categories:

Event Category	Example
Account status	Average bank account balance trending down (by X standard deviations)
Behavior-Account	Roaming charges incurred or within x% of the limit
Behavior-Device	Failure of device/machine
Behavior-Person	Digital browsing – showing purchase interest/intent
Calendar	Major shopping holiday approaching
Contract-Account	Changes in the account/contract terms & conditions
Environmental	Severe weather alert – hurricane warning
Forecast	Model score updated – Churn/attrition score rises above a threshold
Inactivity	No activity (of a certain type) in the last 30 days – e.g., no deposits
Law/regulation	Change in the overall privacy policy
Milestone	Birthday – Age changes (milestones such as 18, 55, 65, etc.)
Product/Service	Replenish – consumable products, such as printer ink
Product-wide/Service-wide	The interest rate on all accounts of type X increases by x%
Profile-Person	Investable assets increase (or decrease) by x% (or crosses a threshold)
Transaction status	Order status change (disruption in availability, timing)

Below you’ll find an inventory of event-based marketing (EBM) and complex event processing (CEP) use cases for customer experience management. In each, the system senses behavior and alerts a user or another system to the unusual activities or conditions that warrant further investigation or action.

Vertical Complex Event Processing Use Cases

Fiserv: Consumer Banking and Credit Cards

Unusual account activity (e.g., large deposits/withdraws)
Unusual account activity trend (e.g., average daily balance down by two standard deviations)
Inactivity pattern (e.g., no transactions in last week)
Missed transaction (e.g., missed direct deposit)
Credit card spend activity use pattern (by spending category)
Insufficient funds pattern
Web or mobile click activity indicating an interest in a product

Insurance

Fraudulent claims activity

Operations

Predictive maintenance systems

Media and Communications

Dropped call pattern or degradation of signal/service
A customer has increased roaming (or other unusual account usages) behavior
Customer in route to a foreign country pattern
Popular programming based on a set-top box and social media insights
Prepaid consumption detection and stimulation
Churn detection

Healthcare

Claims fraud
Care interruption pattern
Fitness monitoring
Hygiene procedures pattern
Healthcare patient monitoring

Horizontal Complex Event Processing Use Cases

Customer Service Center / Retention Department / Loyalty

Customer struggling to get help pattern
Payment due
Strange returns activity
Customer likely wants to cancel service
Customer’s birthday
Customer’s service anniversary (e.g., been a customer for X years)

Marketing / Cross-sell & Up-sell

Customer online interest in a product or service
Customer in store interest in a product or service
Customer in the proximity of a store
Customer usage stimulation – Drop off in use of a product/service
Increase in use of a product/service
Loyalty Program – Monitoring points activity
Loyalty Program – Monitoring points expiration date
Monitoring social sentiment
Monitoring social influencer

Non-CX use cases

Here are some examples that are not for CX, but instead to improve business and operations efficiency:

Algorithmic stock trading such as if Stock A rises by X% and Stock B doesn’t automatically buy Stock B

Transportation security and fraud detection such as an id card used twice in a short time frame (e.g., piggybacking) or high volume transactions on a new account – and then automatically alerting the right parties

Detecting transportation congestion and incidents, and proactive notification of alternative routes

Inferred detections suggesting that a vehicle has crashed (and severity of crash), such as when an airbag has deployed

Communications security such as false alarms going off in a certain time window, not followed by other alarms that would be expected (false positive alarming)

Communications security such as network monitoring for detecting denial of service attacks, and alerting the right parties of this situation

February 17, 2017

3 Tips to Drive Business Value with CX – Fortified with REAL AI

With a first name of “Artificial,” AI has certainly entertained us with its virtual possibilities. Stories of wholesale disruption by robots and fully automated lives make for good movie material, but as of yet, AI hasn’t dominated the marketplace, consumer experiences, or business applications in a monumental way. AI has the potential to change our daily lives, yet for most, its impact so far has been nominal.

As a businessperson concerned with driving better customer engagement, you’re no doubt interested in this topic, yet probably carry some healthy skepticism about the potential for return from your AI investments, and the risk of them failing.

Congratulations! Your suspicion is not only natural, it’s warranted. Here are three tips for how to maximize value from your AI investments, and minimize any risk of disillusionment.

1. Provide predictions about Customer Intent

No doubt, you have scores of business intelligence systems that compile and codify data. They provide customer profiles, program dashboards, and other scorecard reporting of historical results. Although informative, these systems aren’t predicting anything. As such, they are rear view mirrors, providing a view of the past, but not anticipating and generating ideas regarding courses of action that may lead to more optimal outcomes.

Any investment in AI aimed at improving customer engagement must include capabilities to predict customer motivation. Why are they calling? Are they already upset? Are they highly likely to be shopping for another provider? What product or service best suits their true needs? How valuable is this customer over their projected lifetime?

Answers to these questions are always guesses, yet pragmatic AI systems today use proven statistical methods to minimize errors in predictions, calibrate themselves with feedback loops, and provide confidence intervals so users understand their range of applicability.

For example, it’s feasible today to have a portal providing your marketing employees with accurate predictions such as:

Customer value
Churn likelihood
Loyalty to brand

For service agents, predictions like:

Customer sentiment
Reason for calling
Nature of problem

For sales personnel:

Price sensitivity
Available budget
Perception of value

Effective AI has to improve your ability to understand what impels your customers to behave the way they do, or the way they may act in the near future. Work backward from these insights, and demand that your AI systems and vendors can prove they have experience extracting insights from available data, and in predicting and surfacing these items.

2. Make dynamic suggestions to better serve the Customer

Consumers do business with brands that provide repeatable value. That value comes from not just positive product use, but also from an enjoyable and smooth buying process, a friendly and efficient on-boarding experience, and stellar service.

As consumers experience a brand during those journeys, they rack up the score, keeping tally of the relevance and effectiveness of the systems and people they encounter along the way.

Any AI system worth its salt should provide ranked suggestions either directly to customers, or to customer facing employees such as:

Next Best Offer: The most relevant product needed, and an individualized incentive on it that will be both compelling, yet still economically affordable to the business.

Next Best Service Action: The best thing an agent can do next to maximize the chance of reaching an effective and efficient solution to the service problem at hand.

Next Best Sales Activity: The best action for a salesperson given available leads, accounts, contacts, and opportunities.

For the marketers responsible for providing next best offers, AI systems should help them recognize buying patterns, automatically perform tests, filter out offers that don’t apply, and statistically rank the best content & promotions for the right individuals. AI should even suggest the best timing for those recommendations.

For service workers, AI should deflect routine service requests to automated or self-service channels, guide agents on complex service cases, surface potential solutions to issues, and help gauge the sentiment of the customer during the process.

For salespeople, AI should predict the best contacts to engage with in an account, the activities most likely to move an opportunity to the next sales stage, and which accounts to spend energy on to maximize close rates and quota attainment.

3. Install a system that learns in Real-Time

Your world changes every day. As a professional, you wake up every day to news of competitive threats, new opportunities, and market conditions that vary the effectiveness of the strategies you employed yesterday.

If you were slow to react, or simply ignored these factors, you’d fully expect your overall business performance to degrade, so you listen carefully to these environmental conditions, and you adjust accordingly.

Think about your AI systems the same way. They must include adaptive mechanisms, where recommendations made are monitored, in real-time, and dispositions are fed back into the machine, so it can learn from its success and mistakes. Marketing, service, and sales systems receive feedback constantly in the form of customers either ignoring your treatments, or responding to them, so ensure your AI system uses them. Your AI system should rapidly improve its performance, as it’s fed more data, and as it tunes itself. If it’s not, after a short trial period, start asking some hard questions to your provider.

Make sure your results (even if delayed), are monitored, measured, and understood. An accurate measurement of the real business value from AI comes when you understand the baseline, and can measure the lift you get when you employ the insights and recommendations delivered by AI.

Track response rates, conversion rates, incremental revenue, return on investment, and compare to what your vendor promised, what you expected, and what you need to achieve.

AI is a broad topic, yet to improve customer engagement and your outcomes, boil it down to these 3 things; understand customer intent, make relevant suggestions, and learn in real-time so your performance improves over time. If you do these, you will realize REAL value from AI.

January 27, 2017July 15, 2018

AI in CX: Real or Superficial Intelligence?

By all accounts, 2017 has ushered in the dawn of the newest Artificial Intelligence (AI) era. Most technology hype cycles follow typical paths, quickly shooting up, often followed predictably thereafter by a meteoric reentry to reality. Typically, the entire flight takes place over a decade or so, as the fuel of inflated hype burns out, and the gravity of commercial application pulls down on its excitement to test its true value.

AI, however, seems different. It has appeared, drew much fanfare, and then disappeared several times already – more akin to a comet, flaring a tail of excitement with each new orbit. As it reemerges, nearing the heat of expectation once again, it lights up with a spectacular plume, flung into space for another long dark hiatus.

AI history suggests five such orbits already – so is it destined for cold dark space soon?

Superficial AI

Regardless of the metaphor du jour, what we must inspect is the true value returned today, not the imagined expectations of tomorrow. The best test of commercial viability is not an intelligence test; it’s whether consumers are getting more value, and if the business offering the products & services are using AI technology as leverage, providing those things with higher margins.

For example, my mobile device is now my phone, my Garmin, my camera, my alarm clock, my digital assistant, my video recorder, my dictation device, my virtual reality device, and so forth. 20 years ago, it might have cost me $5,000 for these services. Today, I get it all for $500 – $700.

We’re all under pressure to do more in the same amount of time. To that end, these devices have become indispensable – they are essential to modern day survival – adapt to them, use them efficiently, or you’re passed by.

Therefore, by some measures and definitions, AI has delivered this time around. Personally, I don’t care when a big company announces their sixth AI acquisition, or what their advertisements or creative animations say. In my view, the proof is if customers are buying, are satisfied with those purchases, and are reporting their lives are easier, more productive, and more enjoyable.

Businesspeople must apply the same tests. Can they deliver better customer experience with AI? Are their product & services measurably smarter and more efficient?

If they aren’t passing those tests, then it’s just superficial AI.

Real AI Value in CX

As we all admire the latest bright tail of inflated expectation, let’s study what AI has really contributed to delivering better customer experience (CX) this time around.

For starters, look again at that magical device, the smartphone. It streams location data, activity levels, browsing preferences, timing behavior, and the like. Businesses consume this contextual data, and use decision hubs infused with AI algorithms that in less than a second calculate a next best action or insight. That’s real! Big banks, telecommunication / technology firms, and retailers are doing this today to improve acquisition, on-boarding, cross selling, and retention rates.

For consumers, the insights automatically delivered include recommended products, drive time estimates, calendar reminders, and service alarms. Alerts & notifications remind when bills are due, when fraud occurs, or when more exercise is required to meet goals. Cars drive & park themselves, thermostats learn, and media services understand consumer preferences. Customers can interact with machines by simply speaking to them.

For the marketers responsible for engagement strategies, AI now recognizes buying patterns, automatically performs A/B and multi-variate tests, which ranks the best content & promotions for the right individuals, and even suggests the best timing for those recommendations. For salespeople, AI predicts the best contacts, opportunities, and accounts to spend energy on to maximize close rates. For service workers, AI deflects simple service requests, and guides agents on complex service processes to improve time to resolution, ultimately improving customer satisfaction.

Simply put, there can be little argument that AI has delivered value during this orbit, much of it in the form of automation as opposed to higher-level intelligence. Fewer marketers deliver more relevant and better-timed tactics. AI assisted sales means higher quality pipeline with sharper close rates. Contact center managers relish shorter handle times and more efficient call resolution with less staff, and consumers enjoy shorter wait times and voice / bot-assisted service. For those using AI, NPS and customer satisfaction scores are on the rise.

All of these outcomes are commercially feasible. Every business (not just the avant-garde) must rapidly incorporate these proven technological capabilities. Hesitate, and the likely result will be eventual irrelevance.

What’s next – In my lifetime?

With all this said it’s back to our question. Can AI keep delivering, or is it bound to let us down soon?

As humans, we love to dream. That’s important. In fact, regardless of how fast machines move forward, it’s still something that separates us from them. We envision a fanciful future, and plot our course toward it. Along the way, we stumble, get humbled, get up, and plot again. This is our nature. Each step along this evolutionary path, we create and refine machines that help us achieve our dreams.

Our vision seems unchanged. We long to make life easier and more enjoyable for more of us. To do this, we must continue to refine our existing tools, and invent new ones that assist us, and make up for our physical and human limitations. No different from our first instruments, modern day smart tools take over tasks we were never very good at, or simply couldn’t do. They help feed us, optimize our resource consumption, and make our very survival possible. We are already dependent on them, and there is no turning back.

This is also true for customer experience tools. Our expectations are high and climbing. We expect to interact with brands that listen, understand our preferences, react accordingly, and when something goes wrong, can turn on a dime and make things right instantly.

When I enter a website, I expect the search to be intelligent, the user experience to be delightful, and the checkout process to be flawless. If I chose to do all this while mobile, I expect the same experience on my smartphone. If I need help, my first reaction is, “why did things go wrong in the first place…how could this have been prevented,” and then I test if resolution comes fast with low effort – and does the business learn from the mishap.

This is the new normal. Unfortunately, many brands today are not delivering on this type of customer experience. The bar is high, but the elevation of game is not so much a demand from technology as from organizational re-tooling and reorganization to accommodate for technologies already commercially available.

Technological advancements will continue to accelerate. Smarts will show up in more devices. We will demand our machines become more human, especially in delivering customer service and better experiences. As humans, we love a personal touch, a social exchange, a sense of community and belonging. So far, machines have not been able to deliver on any of these aspects. That’s changing.

Presently, there is very interesting research going on to bring more human-like aspects to machine interactions. Google’s DeepMind research lab has made impressive gains in speech synthesis (text-to-speech) in a project known as “WaveNet” where robotic voices are becoming a lot less robotic. Similar advances in Chabot research is leading to smarter bots able to remember details, learn right from wrong answers, and hold basic conversations. You can try one of the better ones at http://www.mitsuku.com/

These developments are exciting. The possibilities are enormous. Yet until these become commercially viable and noticeably better with true customer engagements, you should train your eyes on what is real in AI today. For now, focus your investments and efforts on delivering real CRM value from AI tech today in the form of things like simple service request deflections, intelligent routing to the right agent, relevant product recommendations & next best offers (based on individual behavior profiles), and guiding salespeople with next best activities.

Meanwhile, keep close tabs on these other AI CX innovations as they progress, take some calculated risks on a few promising areas, and prepare for the next revolution of AI. The AI comet will be back shortly.

December 3, 2016February 6, 2017

Machine Learning Measurement: Episode #5

In this 5th short video in my Machine Marketing Series, I explain Machine Learning Measurement and monitoring techniques.

I cover the concept of a Model Factory and Dashboard where you can:

Measure Machine Learning Lift employed in Marketing and CRM programs
Monitor Model Features and Model Data use
Pit models built for Customer Experience (CX) against each other (Champion / Challenger)
View KPIs (Key Performance Indicators) on a business Dashboard

Use these tips & best practices to benchmark against your efforts, and compare how your platform stacks up in using machine learning in your marketing and Customer Experience programs.

November 28, 2016February 6, 2017

Marketing Results: Machine Learning Episode #4

Getting positive marketing results for your customer engagement efforts is what it’s all about.

In this episode, I explore what to expect in terms of outcomes when you effectively employ machine learning (ML) and artificial intelligence (AI) in your marketing efforts.

I cover performance ranges I’ve seen actual companies achieve including:

Churn & Attrition reduction
Response Rates for your marketing execution & tactics
Sales Lift increases which means revenue to the bottom line
ROMI – Return on Marketing Investment
NPS – Net Promoter Score improvements
AHT – Average Handle Time reduction

You can use these figures to benchmark against your efforts, or even help you build a business case for embarking on using machine learning in your marketing and Customer Experience programs.

November 9, 2016May 6, 2020

Machine Learning & AI Use Cases for CRM

Updated: May 6, 2020

Below you’ll find a good organization of classes of AI use cases for CRM (Customer Relationship Management) & CX (Customer Experience). Presently, there are countless examples of applying Artificial Intelligence (AI) and Machine Learning (ML) to solve a broad class of problems beyond CX, from cancer diagnosis, robots in manufacturing operations, streamlining product development, detecting anti-money laundering, spam filtering, improving cybersecurity, to self-driving cars.

Great marketers and CX experts have always been change agents. As new-age change agents, they focus on AI use cases that enhance CRM, and have either already been proven to lead to successful outcomes, or that show significant commercial promise.

But what is AI and how can it be used to take customer engagement to new levels? To make it simple, think of AI as the application of software technologies using standard hardware for:

CRM AI purposes –

Task automation:
- Doing something in less time with fewer steps
- Accomplishing something with less total effort
- Doing something with less or even zero human intervention
Detecting, classification, and alerting:
- Sensing and understanding a current problem (customer is dissatisfied or struggling)
- Sensing current customer intent (wants help; wants more information; trying to buy)
- Placing similar things together (customers, products, content)
- Informing someone or some other system once a certain confidence threshold is reached
Predicting something is likely to happen:
- Customer likely to buy more (or less; or never buy again)
- Customer is likely to call for service
- Contract is likely to close
Suggesting (in reaction or pro-action) a course of action leading to more optimal outcomes:
- Doing X will resolve the problem
- Recommending Z will satisfy the requirement
- Offering Y will increase customer lifetime value and/or maximize profit

Machines are capable of doing these in a faster, more accurate, and more cost-effective way versus humans doing them manually. When that happens, the result is a superior customer experience and better business results.

CRM is a technology that improves customer experience (either unassisted or assisted by an employee) as it relates to customer’s interactions with a brand’s marketing, sales, and service, and the fulfillment of commitments by those three functions.

Artificial Intelligence (AI) technologies –

Regarding different sub-areas of Artificial Intelligence and its class of technologies, consider this organization:

(Note: these categories are not mutually exclusive in that a given application may benefit from one or more of these technologies)

Machine learning – Algorithms using statistical methods designed to predict or forecast something. The learning is either supervised (i.e., assisted) because we either know the outcome trying to be achieved (or we have a teacher to inform us) and can determine if the model is right or wrong – or unsupervised (i.e., unassisted) because we have no idea of the patterns or outcomes that may be best). The system “learns” if over time its predictions and performance improve.

Text, Video, Audio, and Image Analytics – this code scans unstructured data, and finds entities/objects, and classifies them (and/or extracts them). Deep learning is an advanced form of image analytics – using a technique with layers of brain-like neurons (Recurrent Neural Networks (RNN) or Convolutional Neural Networks (CNN)).

Complex Event (Monitoring) Processing (CEP) – an engine fed streaming data from one or more source, which detects certain patterns and then initiates follow on actions or processes.

Deep Q&A – using an AI system that has access to previously compiled knowledge sources; it scans, filters, and presents the best likely answers to questions.

Natural Language Processing – this code translates spoken voice to text, or to some other form useful as input to another system. It can also do the reverse, translating some other system’s output to spoken voice. It also can translate from one language into another, or just detect the language.

Natural Language Generation – this code takes input (images (still or moving), text, etc..), and generates text descriptions of what its seen or found.

Numeric Analytics – this code uses commonplace mathematics (simple formulas) to surface insights, focusing the receipting’s attention to aid their future decisions.

Robotic Process Automation – this code repeats tasks that it’s instructed to repeat.

Deterministic rules – if/then/else statements that don’t change unless a programmer changes the rule.

Click here for an AI diagram of how I’ve organized these into building blocks so you better understand where each belongs in terms of its contribution to automation or true machine intelligence

CRM AI Vehicles –

Regarding implementing AI use cases in CRM for improving customer engagement (e.g., using any code and a combination of the above algorithms), consider these main vehicles:

Virtual Assistants – a module designed to aid and assist a human with decision making, problem resolution or task completion including:
- Chatbots (such as Facebook Messenger)
- Conversational interfaces (such as Alexa, Google Home)
- Office assistants built into desktop software or mobile apps (e.g., automatically schedule meetings)

Digital Recommendation Engines – ranks and bundles content recommendations, product recommendations, offer recommendations, and service recommendations and uses containers/spots on web, mobile, agent desktops, email, kiosk, and other digital devices to serve these to:
- An employee or agent to sell a product or solve a service case
- Engage directly with a consumer through any digital device

CRM AI Use Case Categories –

Given these areas, there are hundreds of potential valuable AI use cases for CRM. Here are some examples:

Marketing

Profiling and Tactic Execution:

Predicting missing or outdated customer data values

~Forecast data value using time series, or simple inflation adjustments
~Predict data value from an image

Segmentation, clustering, targeting
Next best product / service recommendation
Next best content recommendation
Next best promotional recommendation
Next best channel to engage on
Next best time to send
Next likely transaction (and timing of it)
Next best search keyword

Predicting Customer Behavior and Value:

Models to predict customer churn
Models to predict customer preferences (including a preference for a particular product/service)
Models to predict customer lifetime value
Models to predict current customer value
Models to predict customer wallet share
Models to predict customer re-activation

Dynamic Price Optimization

Optimize retail price of products/services
Optimize incentive levels (discounts) of products/services
Optimize price paid for ad buys

Marketing Operations / MRM / Marketing Planning:

Tagging/filtering content – Recognize an object in image or video
Tagging/filtering content – Match an audio track to known copyright material
Tagging/filtering content – Classify image content type
Content generation
View trending topics
Forecast market size, marketing revenue/costs
Budget allocation planning/media mix optimization

Influencer Marketing:

Content identification

Sales Automation

Next best lead
Next best sales activity
Next best sale rep
Sales contracts
Sales and demand forecasting
Competitive intelligence

Customer Service

Customer Service / Support

Next best service, support, training action to resolve a complaint/problem
Intelligent routing: Find the next best expert/agent to resolve a complaint/problem
Text and images uploaded to support resolving cases
Tag and classify support tickets
Support process automation
Estimating service volumes, wait times
Optimize scheduling and support resource utilization
Contact center volume forecasting
Estimate wait time
Optimize scheduling and support utilization optimization

Predicting Customer Service Behavior:

Models to predict levels of customer frustration or satisfaction
Sentiment analysis with all forms of data.
Reasons behind high churn likelihood
Models to predict intent (likelihood of a customer calling and why)
View trending support topics

November 4, 2016February 6, 2017

Smarter CRM: Machine Learning Episode #2

You want Smarter CRM, don’t you?

In this “Mini-Cast” on machine learning (ML) and artificial intelligence (AI) for marketing & CRM, I outline the numerous applications you can use to improve customer experience.

In this second episode, I delve into the various areas of AI & ML applied to Marketing, Sales Automation, and Customer Service where ML & AI play huge roles in taking those functions to new levels of insight and intelligence, enhancing productivity, effectiveness, and delivering better customer engagement.

Episode #2 on YouTube: