(This is a rework of an article I wrote a few years back, with a bent toward Data Science)

Being effective as a data scientist means paying attention to the things that matter. That recipe includes metrics, algorithms, modeling — important ingredients, all — but it also means using your brain to pay attention, as a human, to what your customer, another human, needs.

Waitresses are great at this. And from what I’ve seen, the over-worked ones at diners and dives often have the best 6th-sense for how to handle a customer. When it comes to your restaurant (your company, your app, your website) and your hungry patrons (your paying customers, co-workers, the marketing team, your boss), and your waitress (uh, YOU!), answer me this:

“Would you tip you?”

Bryan Eisenberg often writes about the “Conversion Trinity” as the effective formula for improving conversion — this dates back from when we were at our consulting agency (back in the day). I feel we can apply the same principle to data science; its three core concepts applied to data science map directly to the same ones that you’d expect from exceptional service experience at a restaurant: Relevance, Value, and Call To Action.

Of course, that’s not the parlance used at a great eating establishment. (“86 Those Words, Chef!”). But the principles are the same:

  • Relevance: From a Data Science standpoint, we talk about about being relevant to the customer’s wants, desires — general goal alignment. Sometimes a client, say a marketing or sales team, is looking explicitly for insight to gin up ideas on how to focus efforts; other times, they are looking for help in choosing between two or more good ideas. Sometimes, hopefully rarely, they’ve committed already and only want to hear about evidence that supports their idea. (hint: show them how getting data science involved earlier would make their ideation more powerful.)

From a waitress’ perspective, if I walk in and ask “what’s good?” a great waitress immediately has something to recommend. And she’s familiar enough with the menu, that if I get a “no-no-on-carbs” look in my eye when she says “pasta” , she immediately  pivots to a steak or salad. If I walk in and say “I’d like a steak”, then a great waitress doesn’t spend anytime singing the praises of the risotto, fine as it may be.

  • Value: From a Value standpoint, we talk about if the customer knows why this is exactly the right sort of data science project that gets them to their goal we identified under Relevance above. Has the value proposition been explained well? Does your customer understand how the proposed data science effort will force-mutiply their own efforts? You, the data scientist, might get it….does Marketing?

From a waitress’ perspective, why is “what’s good”, well, good? Maybe it’s “we’re known for this throughout the South”, or “we’ve been making it the same way since 1912” or “we use special mushrooms gathered under the light of a full moon”.

Or perhaps we’re Value in the literal sense: “we serve Prime cuts you can’t get anywhere else, and we do it for prices that make our owners weep”. However the waitress sells it, a great waitress implicitly knows the Value has to match the Relevance.

[Personally, this is why I immediately tune out a waitress who tells me she’s never had dish X, but “has been told” it’s really excellent. This is like the Vegan animal rights activist who wears a leather belt. Just say “I’ve never had it, but I get the portobello burger all the time  and I love it!”

I don’t have to like the same stuff as you but I want authenticity, not a blog aggregator of what others have said. To that end, I want a waitress who likes to eat! Never trust a skinny waitress! Ah, but I digress…]

  • Call to Action: From a Action standpoint, we talk about whether it’s obvious what to do next, and whether the customer has the confidence to take the next action. If the data science project results in a predictive model, how are we going to apply it? Do we have the resources to apply it? Does it get punted over to Engineers to be implemented? Or are we going to use the results to do a quick field test to confirm the results and then apply a full-on sales blitz?

From a waitresses perspective, this means helping me narrow down my choices (“yes, we have half a dozen different kinds of pasta, but the only two to consider are…”), addressing concerns I might have (“this is a great dish except if you’re allergic to peanuts”), and then asking for the sale (“Can I bring you that salad you have your eye on? You can always get something else if you’re still hungry after”). In most cases, she needn’t even ask for the sale at all, since the customer will sell himself (with her help, ‘natch).

If you get Relevance and Value correct, you’ll have to go out of your way to goof up Call to Action — and even if you do, it’s easy to spot and correct later. But goof up the Relevance/Value marriage, and the best Call to Action in the world will sound hollow and forced. The waitress does the same thing by establishing rapport with you a.s.a.p. to help you get to Action.

And don’t ask if I want to see the dessert tray: Just bring the damn thing over, and while we ooh and aah, then ask “may I entice you with a dessert?” That’s a damn fine combination of Call to Action (“Yes”) and up-sell through creation of new Relevance and Scent.

Think about this the next time you’re at a restaurant and you get great (or horrible) service. You’ll soon see the parallels. It’s easy to tip a great waitress, because she is a joy to work with.

Shouldn’t your team’s integration with data scientists be this way? All those metrics and analytics and models … it’s nothing more than supporting materials  to help your business consider what it’s doing from a goal perspective of Relevance, Value, and Call To Action.

Given how you currently handle this at your company, would you tip you?

Would You Tip You? (Data Science Edition)

Post navigation

Leave a Reply

Your email address will not be published. Required fields are marked *