skip to main content

Promotional Prestidigitation

Acxiom Last Updated July 5th, 2017
Promotional Prestidigitation

Last month I showed you a magic trick where a magician takes an ordinary deck of cards (representing your customer file) and with several flourishes produces a marketing and analytics strategy from thin air. If you missed the trick, it’s called “counting cards,” and here’s the link again.

Having done magic throughout my misspent youth and now working in analytics at Acxiom <insert your own joke here>, I can predict with great probability how people will react to magic tricks. Some will tell you they know how you did it, some will say they’ve seen it before, some will ask, “Who is this annoying pathetic magician?,” and some will say, “Show me another magic trick.” (Magicians like to hear that.)

Let’s discuss the first two reactions, and then I’ll show you another trick! Yes, most companies are doing a version of “counting cards,” but when you ask them to show you their trick it’s clear they are not doing it right. Some versions are unstructured… we want more of her, more out of her and maybe a younger her. Some versions include math… we have a 34 percent attrition rate, we are seeing 17 percent fewer new customers, and so on.  I especially see this with the monthly analysis of online/digital exposure data… more opens, fewer clicks, more impressions, more bounces, less time on the site, etc.

But few will tell you why it’s happening, the actions required to change customer behaviors, how they plan to acquire more high-value customers and the value associated with all those actions. Very few know how to analyze that data, develop meaningful insights and make prioritized marketing/analytics recommendations that are sequenced by their bottom-line impact.

OK, thanks. You’ve listened to my patter patiently, so now let’s do another trick. The cool thing about this trick is you can use it for retention, cross-sell and acquisition – it plays well across all use cases. In general, analytics are just tools (scores), so the real value of the tool comes in how you take action, how you measure success and how you develop repeatable, scaled processes. For example, it would be naive to think an attrition model alone will solve your attrition problems. <audience murmurs here>

This trick is called “promotional prestidigitation.” I’m hearing different names for this trick in the marketplace – price sensitivity, promotional depth, markdown optimization, promotional cadence, etc. Most companies do this trick by “gut feel.”

OK, here we go, I’m rolling my sleeves up now, and you see there’s nothing up there. Just like we did with “counting cards,” think of your customer file as a deck of cards. I’m going to fan the cards out – the kind of table fan you might see from a blackjack dealer in Vegas. On the ends of the fan we see our extremes – the full-price-only shopper on one end and the discount-only shopper on the other end. Every customer falls somewhere in between. This trick is about placing each card (each customer) on the fan. To what extent does a customer buy on promotion?

To determine this we’re going to look at a variety of factors for each card:

  1. Promotional usage. Did a customer raise his hand to receive a promotion? Did he “do something” like use a coupon, enter a promotional code during checkout or take advantage of any special offers (online or offline)?
  2. What is the timing of a customer’s usage, to what extent did the customer shop during a promotional period relative to non-promotional periods? Does the customer buy at a higher rate during a promotional period?
  3. Where do customers buy within a product category, are they buying on the low end, middle or high end of the category? Regardless of a promotion, do they seem to buy on the low end?
  4. Share of Dollar. For every dollar spent, what percentage was spent on promotion?
  5. To what extent does a customer wait for a deeper discount (e.g. 40 percent vs. 20 percent)? Are they high, medium or low in their promotional depth?

There are some cool variations of this trick where you can look at these five factors alone or in combination (a simple composite score). Some magicians like to develop a probabilistic model predicting where you fall on the fan; others prefer a simple index.

Either way, a good magician knows that deriving the score means nothing without an activation plan. By the way, never drink and derive when you do this trick.

OK, let’s pause here. We now have, for every customer, a promotional (price) sensitivity score driven by past behaviors. Some of you may be asking how this will help with a new customer (where you have very little past behavior).  I did say earlier this trick could be used for acquisition as well as cross-sell and retention. The quick answer is we need to use segmentation for new customers, another tool and another trick for another blog (tune in next month). But for now, know that a consumer’s price sensitivity plays well with segmentation.

As you can imagine, different segments have different price sensitivities. For new customers, where you have few if any behaviors, we can ascribe the price sensitivity of their segment to them in the short term and as they reveal more behaviors over time we can replace our segment level insights with an individual-level predicative score.

OK, let’s pause again. Most magic tricks are easy, once you know the secret. The skill comes in how you present the trick. There are very few bad magic tricks, just bad magicians.  It’s important to prioritize your strategic objectives. What is the goal of your promotion? Are you trying to drive upper funnel (new customer) acquisition to your site or store/branch? Are you trying to re-engage a lapsed customer? Are you trying to move decreasing-value customers up the value chain? Maybe you’re finding that many of your customers are migrating to lower-value segments and you want to stop (or slow) that migration.

To prioritize your strategies, practice last month’s trick, “counting cards.”  This trick will very clearly reveal audiences that need the most attention with the highest dollar impacts.

Let’s do one use case together … attrition (a big problem with many of my clients). In some industries, attrition happens suddenly, like when a customer cancels his subscription. In other industries, attrition can happen slowly, customers slowly growing away from your brand. Customers have their own cadence. Some shop monthly, others shop seasonally, others really only shop once or twice a year (especially in luxury retail). If a monthly customer goes dark for three months, that’s very different than when a seasonal shopper goes dark for three months. One size does not fit all with attrition. By the way, modeling for attrition is another trick, for another blog (stay tuned). But the point here is the former customer warrants a re-engagement campaign with a personalized offer while the latter customer does not!

So STOP sending the same emails, at the same cadence, with the same message, and same promotional offer to customers who go dark for 60 or 90 days. Use segmentation to personalize the message, use your promotional score to drive cadence and personalize their promotional offer. For customers with high promotional scores (high price sensitivity), target them to clear excess inventory. For customers with low promotional scores, STOP offering them big discounts.

Let’s wrap up this trick with the importance of measurement. For every dollar you give away on promotions, you must be able to measure the incremental sales lift. Do A/B testing and experiment with different creative, use test/control to measure incremental sales lift and return on ad spend. Continuous experimentation and measurement will lead to optimization. And invest in automation to make all the handoffs fast and easy. Do this in an open garden environment if you want to optimize, and don’t let your partners grade their own homework.

It’s easy to get lost in “big data,” easy to get misdirected. That’s what magic is all about, misdirection. I believe amazing things are happening in your customer data – every single day. The problem is we don’t always pay attention, or we pay attention to the wrong things. Becoming a good magician is about paying attention to the right things. Try these tricks, and practice!  If you need help, reach out to Acxiom. We can help.