BrainTrust Query: ‘Great’ Can Be the Enemy of ‘Good’ Using Customer Data
Through a special arrangement, presented here for discussion is a summary of a current article from the Mark Heckman Consulting blog.
When I talk to brand people, especially the ones who are relatively new to the food industry, one of the first questions they have pertaining to retailers is, "Why are so many doing so little with so much customer data?"
Scars from experience have led me to the painful conclusion that "great is indeed the enemy of good" when it comes to retailer customer strategies. Most that are offered today are too complex to execute and maintain. These complex strategies require content and funding that can only come from the buckets of dollars retailers use for mass marketing programs.
In fact, many retailers that have assembled fairly impressive customer databases over the years still struggle to leverage these data efficiently and effectively. I count myself among this group. Trying to ingest all this data can be tantamount to "drinking out of a fire hose."
To be blunt, retailers and brands remain all too entrenched in their more traditional programs and their short term benefit to divert funding to targeted customer specific initiatives. These programs are still sacrosanct to most retailers and brands, that enjoy the predictability and reach these programs afford.
So that leaves us with starting simple, building incremental credibility, and subtly diverting dollars and human capital to a more targeted marketing approach.
Tracking this discussion to its logical conclusion, retailers that are struggling with their data need a "bridge" plan, one that gets them in the game, but does not tax the entire enterprise by competing with mass marketing programs for funding and attention.
Step One: Identify your best shopper base. Track them, reward them, and make certain best shoppers know that you know they are special.
Step Two: Devise a way to recognize first time customers. If you have a loyalty program, develop a mechanized bounce back reward or message at the point of the very first engagement that incentivizes a second trip. If there is a second engagement, recognize that as well. If, after a reasonable period of time there is no second engagement, devise a plan for that scenario.
Step Three: Track defection and attrition. Devise intelligent ways to understand how to allocate resources to ameliorate losing shoppers or share of shoppers’ wallets.
Focusing on just these three aspects of Customer Relationship Management will enable retailers to smartly use their customer database in a measured, reasonably affordable way. Once "wins" are gained in these three areas, more specific strategies can be employed, hopefully with funding from brands and other sources that are convinced your CRM program works and delivers results.
Discussion Questions
Discussion Questions: Why are many food retailers, despite having invested in building significant customer databases, still not fully leveraging their customer data? What are some minimal steps retailers should be taking to capitalize on the available data?
I wonder if it’s because they think of it as a database rather than a source of answers. Mark’s starting points are pretty good, but reflect only a small piece of what they can learn from that database. Retailers and their vendors should be looking at assortment, category co-purchasing, frequency of shopping v. norms, and the like to improve their business. I’m betting one good analyst working for the retailer could generate enough good information to more than pay their salary. And a good IT department can build automated rules that incorporate this learning for identifying ongoing marketing opportunities.
Massaging data takes money, and until a food retailer can be shown a significant return on investment, most are hesitant to allocate the necessary funds. The article details three steps that all retailers should take.
I suspect many food retailers who have built large databases and are not using them had poor strategic reasons for the database in the first place. Leveraging a database requires having goals that intersect with shopper motivations.
Mark’s recommendations are excellent. If they fit with the retailer’s vision, they will work well. If the retailer has another vision, the database may still prove a valuable resource. But it must be thought through in terms of corporate goals, shopper attitudes and of course, the ability of the retailer to access the database as desired. Having the right talent on board will make the difference.
One of the challenges is the continual stranglehold IT departments have on their organizations. One way to begin to leverage this data is to allow access to outsourced resources to leverage and activate this data into actionable programs. All too often internal politics and a sense of urgency simply mires these ideas and concepts. Outside resources are hungry for the business and will stimulate activity and response. Listen and talk with your best customers and the staff that deals with them on a daily basis. The concepts and things you can do based upon these discussions are invaluable. Designing and activating just one cross-channel solution based upon existing data will convince your customer that you’re at least listening. Perhaps the executives will begin to understand that this is not a luxury, it is a must do!
I agree with Adrian. It’s time for third parties to step in and offer Saas solutions for all this big data. The analyzing of the data itself does not need to be proprietary. It’s what you decide to do with it that will need the retailers brand spin.
Food retailers clearly aren’t leveraging that ocean of data. One key benchmark is “do you know your best customers, and if one of them leaves you, can you react?” The answer is nearly always “no.” Given the fact that the lifetime value of a good customer is in the tens of thousands of dollars, you’d think they’d like to be able to execute on that strategy.
As usual, a very thoughtful piece from Mark!
It reminds me of a different Mark (Mark Zuckerberg)’s letter that accompanied Facebook’s filing for its IPO, in which he says:
‘We have the words “Done is better than perfect” painted on our walls to remind ourselves to always keep shipping.’
All too often the data mining, data warehousing, BI, and analytics tools used to handle enormous volumes of retail data are designed to address every possible question, every possible scenario, and every need of every business stakeholder. As a result, they are complex and difficult to use, incredibly resource-intensive to configure and deploy, or both. So, implementation projects take months (or longer), and the end result is a system that *can* do everything, but ends up doing very little. A 100% solution usable by 5% of people.
We strongly believe that is the wrong approach. We aim to build analytic software that is stunningly simple and intuitive to use. To do that, we only attempt to handle the most important 80% of cases, which represent 95% of what people really need to do on a day-in, day-out basis. A 95% solution usable by 100% of people.
Of course, there are other enemies of the “good”: competing priorities, urgent IT fire-drills, internal politics, from-the-gut decision-making, etc. But it’s always a shame when a retailer can align itself behind a vision for truly understanding their shopper, and then they fail because they bite off more than they can chew.
Other than the money, self-service retailers are NOT shopper facing. Their businesses are driven by stores, logistics and the supply chain, not by the minutiae of individual shoppers. Even programs like dunnhumby can become complex structures to leverage money from their suppliers.
Mark’s suggestions are eminently reasonable because of their FOCUS on a very few specific actions. But to apply a quote from one of the top three global retailers, “Why would we do this if no one is paying us to?”
Understanding and effectively utilizing the massive amounts of data coming from multiple sources requires a retailer to create a well-defined plan including goals, methodology, required data sources, and sound analytics infrastructure (hardware and software.) And I agree with Dr. Needel in that a retailer will need to look at multiple data sources in order to get a complete understanding of the business.
Too often, retailers have started an analytics initiative without a complete roadmap. I’ve included a passage below from Alice’s Adventures in Wonderland that puts into context an analytics initiative without a plan.
“Would you tell me, please, which way I ought to go from here?” “That depends a good deal on where you want to get to,” said the Cat. “I don’t much care where” said Alice.”Then it doesn’t matter which way you go,” said the Cat.
Really like and agree with Ben Sprecher’s response.
A few other observations from me:
1. I see retailers (and manufacturers for that matter) trying to apply traditional marketing and research approaches to the new world of customer data. They don’t seem to latch on to the fact that you no longer need proxys, a database can cope with multiple segmentations/dimensions, segments really can exist at a household/customer level, insights don’t have to exist at just one point in time etc., etc.
2. Being great at IT is not the same as being great at data. To build great data insight engines for the business is not the same as best-in-class technology. In fact the two are quite often at odds. The real benefits come from the creation and management of the new data (a customer insight layer) not just storage of and access to the raw data generated by business operations.
3. Most outsourced database options for retailers are too similar to internal IT solutions and take an IT and SLA-oriented approach instead of being focused and indeed measured on the outcomes from using the data.
There are many factors at work here. 1) Cut through the noise by investing in a good analytics tool to perform some predictive modeling and other scenarios to derive actionable insights from the flood of data — 80% of which is unstructured. There are plenty of great tools available today that are simply underutilized by retailers. 2) Examine the retailer’s culture and strategy. tactically, if a food retailers has a loyalty program in place, I’d be willing to wager that it is far more a “frequent-shopper-discount” program than a true loyalty program. Most retailers are giving unwarranted discounts to shoppers whom would’ve bought the products regardless of the discount. How many times are you surprised at the money you saved at the POS after the club card discounts? You would’ve purchased most of it anyway. The overall strategy needs to be addressed and develop a true loyalty program that rewards the behavior that is desirable to the retailer.
In our experience in database marketing consulting, one of the great challenges for food retailers is the low margin level of their products. As a result, it is difficult for many direct-to-consumer programs to pay out in incremental margin and justify their expense.
The best practices in grocery retailing however suggest that database marketing can be highly successful when integrated into the operations of the business. For example, Tesco in the UK and Kroger in the US have built business with dramatic growth by using the data in their customer databases. The data is used to improve product assortment, shelf arrangement, promotional strategies as well as cash register and e-mail marketing programs.
Information derived from customer data can be transformative to the organization; however, the challenges of traditional business silos of buyers, store operations and marketing prevent many grocery retailers from achieving those results. What happens is that customer data is seen as simply a marketing overlay, and fails to justify the significant investment required to build and maintain such a program.
If grocers would knock off these 3 basic items, they would be ahead of the game. I have talked with several who query about more complex promotions based on time of day or SKU, but then execute nothing of the sort as they succumb to decision paralysis.
The substitute is the coupon or merchandising solution that is familiar and comfortable to execute. I agree with the author here. Start with small steps, measure results, then determine if any of the key customer groups targeted would respond incrementally better to a more complex rules set.
Better to do something than to be frozen at the whiteboard.