Focusing on the Customer is Way Easier with AI
An ancient strategy framework that resonates with me segments successful companies into three categories: those that pursue operational excellence (e.g. Walmart); those that push the boundaries with their innovative products (e.g. Apple); and the ones that foster customer intimacy (e.g. Home Depot). Your company doesn’t need to be extraordinary in all three - in fact, trying for all three is likely to produce mediocrity; instead, pick one, align culture and processes around it, and aim to do an average job on the other two.
AI can help create advantages in all three of these; I am quite passionate about customer intimacy, so let's talk about it today. But first, let's define it:
Understanding who the best customers are, and thus choosing to say “no” to some prospects;
Valuing the relationship with the customer above all, and thus allowing some transactions to be less profitable than others;
Giving employees who work directly with customers more control, and thus choosing to have lower consistency and in some cases, even lower efficiency.
Let’s look at each of these aspects in detail, and see how AI can help with each.
1. Understanding the Customer
Every consumer company out there tries to state who their customer is. A view of the addressable market is a required slide on most fund-raising decks. Many try to do customer segmentation and user personas. Most of these exercises produce decks that collect dust. In my mind, a good litmus test of excellence here is the ability to say who their product is a bad fit for. For example:
for Home Depot, a US city dweller who grew up without ever using a screwdriver is not the best fit;
for Tesla, someone living in a rural area with extremely cold weather won’t get a great experience;
for Instacart, a family on food stamps that’s trying to save every penny will not get enough value from service.
This clear-eyed segmentation helps clearly frame the addressable market, and which customers are likely to be most profitable: unsurprisingly, those that gain the most value from the products are likely to be the most valuable ones to the business.
What happens if a company tries to sell to everyone instead? They will:
Acquire customers who will ultimately become detractors, as the product is a bad fit;
Lose focus, as widely differing customer groups will pull their roadmap in opposing directions;
Waste marketing dollars, onboarding low-LTV customers and blasting big audiences with poorly resonating messages.
AI offers capabilities to understand customers a lot more deeply, at scale. From models that determine who should get the next coupon, to identifying top customers when they call in and placing them in front of the queue, to automatically detecting potential abuse - AI lets you go 1-to-1 in your customer interactions. Segmentation becomes a repeatable exercise with low marginal cost; segments become smaller; actions across segments become more differentiated.
Let’s look at some use cases.
1. eBay identified a segment of their ultra-high-value customers that haven’t purchased anything recently. They then had their customer service reps call these customers (extremely expensive action!) to ask if there’s anything they can do to earn their business back. Surprise-surprise, these customers were so shocked by the personal attention (nobody else does that!) that they, en masse, came back. I suspect that today, this outbound “diving save” is an automated, AI-triggered activity that eBay regularly executes.
2. AI provides an opportunity to know the customer even better than they know themselves. Grubhub, a food delivery giant where I worked, had a unique dataset that showed customer satisfaction with restaurant dishes following each delivery. Restaurant owners themselves never had that data! Grubhub account managers, powered by back-office AI tools, would be in a position to tell a restaurant, “Your fries are soggy on Thursday afternoons. Consider re-training the person that is on that station during that shift.“ This kind of insight is, of course, extremely valuable.
2. Valuing the Relationship
Driven by the peer pressure of Amazon, the most customer-centric company, most execs like to speak about how everything they do is about the customer. Words are cheap, especially those that are difficult to quantify. Let’s see if we can devise a litmus test to see if a company truly values the relationship with the customer - not the immediate profit from a transaction.
Here are a few angles:
How deeply does the company understand customer lifetime value, and how good is it at predicting it via proxy metrics?
How much has the company’s reorder rate improved in the past year? What portion of efforts has been directed toward improving the reorder rate? Not margin, not profit, not AOV, not operational efficiency; reorder rate.
How much focus is there on revenue in company meetings, versus input metrics that are all about customer experience?
Is customer service a discipline that’s seen as a cost center, or as a competitive advantage?
Let’s look at some examples.
Have you ever wondered how Amazon can deliver a $5 cord to you, overnight, with free delivery? The cable might just cost them a buck, but delivering it to you this quickly has got to cost them more… they must be losing money on this!... That’s right, they are. They choose to lose money on this transaction, but you’ll feel that your Prime membership is well worth it, and will keep buying other stuff on Amazon. And that’s how they make up the losses tenfold.
Hungryroot is the personalized grocery delivery company where I work. Over the past 2 years, we improved customer retention by 50% using AI; a key lever that AI employs is giving more value to our customers. Our customers get a box each week; we’ve discovered segments where giving more value in each box - literally, reducing our margin on the order - is so profoundly impactful on the customer experience that the improvement in the reorder rate more than makes up for the loss in margin.
3. Elevating Customer-Facing Functions
Which departments deal with customers every day? These groups have the highest chance of both knowing the customer and having the most sway in terms of the potential to turn an upset customer into an advocate.
The level of power that your company gives to those departments is an excellent gauge of customer centricity. Why? Because pushing down decision-making power brings inconsistency and often imperfection. It means some decisions, for example, a too-generous refund, that might not be in the best immediate interest of the company. And yet, the power balance tilted towards customer-facing functions creates a culture where the relationship with the customer indeed can be the focus.
Let’s look at some examples.
Delta Airlines. There are so many Twitter threads about flight attendants giving out 5-20k miles (which roughly map to $50-200) to customers across a whole slew of situations that have nothing to do with the fault of the airline. Their seatmates spilled a drink on them? Give them miles. They helped a fellow passenger navigate the glitchy internet? Give them miles. Each of these is an example of creating fans and memorable moments. Delta knows that true loyalty comes from that.
Chewy developed a cult following after doing something that short-term thinkers would have a heart attack over: fully refunding customers whose pets died and sending them flowers with a heartfelt note. "But the potential for abuse!..", a short-sighted accountant might proclaim, adding, "They don’t even have a pet anymore, it died, they won’t even be our customer!.. " And yet, tremendous (but hard to measure) brand benefits.
The common thread is clear: allow customer-facing staff to treat the customer well - even if, sometimes, the customer is in the wrong.
AI can take this strength even further. It’s well-known that responding to customer inquiries more quickly is extremely valuable. AI can do that for you. AI can also handle a big chunk of the simplest inbound issues by itself, without involving humans. For the remainder, AI can give advice, templates, and tools that your customer service staff can use. Again, the theme is clear: approaching ChatGPT / generative AI singularly as a cost-reduction tool for customer service is probably not a sign of a customer-centric company.
Leap of Faith to Do the Right Thing
It’s easy to rationalize all of these examples as frivolous; as quasi-marketing stunts that produce, at most, press coverage and social media chatter. These customer-centric approaches offer hard-to-measure outcomes, and as such, are prime candidates for cuts when the first bumps on the road appear. And yet, the relationship with the customer is built on moments like this. Focus on short-term profitability only, and you first arrive at mediocrity; the final destination is the dustbin:
Circuit City fired the highest-paid salespeople to get out of the red. Good idea for cost control, right? Well… Customer service suffered, sales tanked. We know how that ended.
Sears reduced investment into maintenance of their stores for cost control. Reasonable idea for a discount store, right? Well… They seem to have taken it too far. End result? Dirty carpets, declining foot traffic, bankruptcy.
Taking this concept even further, the best have the humility to understand that it is all too natural for techies to build the product for themselves, not for the customer. Nobody is immune; that fallacy results in large commercial flops like Amazon Alexa devices and Facebook Virtual Reality experiences. If only the nerds building those stopped building what they themselves would value, and truly listened to suburban busy parents that are far from tech.
What’s necessary is a bit of a leap of faith: that doing right by the customer will pay for itself in the long run. It’s easy to talk about being customer-centric - up until the moment when you need to give something up to be customer-centric. Strategy, as Steve Jobs famously said, is all about saying “no” to something valuable, giving something awesome up to make room for something better. A customer-centric strategy is about giving up short-term profits, efficiency, and control in pursuit of a long-term relationship.