Is transactional data the key to understanding retail customer behavior?

Discussion
Photo: RetailWire
Nov 09, 2022

Presented here for discussion is a summary of a current article published with permission from Knowledge@Wharton, the online research and business analysis journal of the Wharton School of the University of Pennsylvania. 

The Customer-Base Audit,” co-authored by Wharton marketing professor Peter Fader, argues that firms cannot make fully informed decisions without first understanding their customers’ buying behavior and the actual health of their customer base.

“The customer base has historically been something of a Wild West, lacking a lot of rigor and accountability,” Prof. Fader said in an interview with the Wharton School Press. “It’s about time it was audited.”

To help firms do that, the book sets out the tools and frameworks to assess their customer base, reassess growth objectives, understand how their customers differ in behavior and value, see how the quality of their customers has changed over time, and more.

The authors noted that most businesses already do customer analysis typically on an as-needed basis. A customer-base audit aims to formalize and impose structure to the analysis process.

Bruce Hardie, marketing professor at London Business School and co-author, said, “It’s a systematic review of the buying behavior of a firm’s customers using data captured by its transaction systems to summarize their actual buying behavior. We’re not interested in who these customers are, not what they think, or their attitude. It’s very much a focus on their actual behavior, and providing that high-level overview for top decision-makers in the organization.”

The audit takes in multiple “lenses” in analyzing the purchasing patterns of new and existing customers over time periods while grouping customers into cohorts to understand their behavior and performance over time.

Prof. Fader said such analysis requires “careful handling, careful analysis and careful inferences from data,” but isn’t “too technical.” He believes such analysis is the first step towards customer centricity and should be part of every marketer’s toolkit. Deeper insights into customer profiles should also elevate marketing’s role in guiding an organization.

“As we say in the book, the customer-base audit is unashamedly descriptive. It’s all just there in the data without you having to go out there and work with other vendors or buy a bunch of reports. It’s pure gold,” said Prof. Fader.

DISCUSSION QUESTIONS: Do you believe a customer-base audit as proposed in the article is a good first step on the journey to customer centricity? What are the benefits and drawbacks of leaning on purchasing data to get more of a real-time view of customers?

Please practice The RetailWire Golden Rule when submitting your comments.
Braintrust
"Transactional is where it should always start, because demographics, psychographics, etc. are all descriptive instead of actuals."
"These descriptive measures belong on the BI dashboards of senior management. They will help bring key trends to the surface."
"Can you really be customer-centric if you build a strategy only on transactions, not accounting for the customer decision making process and why they choose this over that?"

Join the Discussion!

32 Comments on "Is transactional data the key to understanding retail customer behavior?"


Sort by:   newest | oldest | most voted
Bob Amster
BrainTrust

The cons of reading the results of such audits is that there can be many factors affecting changing trends. Some of these are external so the retailer may not even know they are impacting the trends. The pros to conducting such an audit is that customer bases do change – for whatever reasons – and knowing that they may be constantly changing is of great value when planning one’s next move.

Dave Bruno
BrainTrust

Without reading more in-depth descriptions of the Customer Base Audit, I am struggling to identify exactly what is new here. The vast majority of our clients analyze “data captured by its transaction systems to summarize their actual buying behavior” and group them into “cohorts” based on those actual behaviors constantly. It’s baked into the business model. Perhaps this model injects more structure and discipline into the process?

Cathy Hotka
BrainTrust

Prof. Fader is correct about our misunderstanding of the customer base, but there is no substitute for boots-on-the-ground observations. Do customers look happy? How do they interact with associates? It’s important to know.

Verlin Youd
BrainTrust

Totally agree. Knowing how customers are behaving prior to checkout is a goldmine. Figuring out how to efficiently get that data has been a great challenge … and some new solutions seem to be helping.

Jenn McMillen
BrainTrust

Transactional is where it should always start, because demographics, psychographics, etc. are all descriptive instead of actuals. You always want to base decision-making on actual data, not inferred. It sounds so old school at this point, but there’s nothing wrong with going back to basics.

Ryan Mathews
BrainTrust

But Jenn, demographics and psychographics inform, if not drive, those actual transactions. We now have tools that give us a much more holistic understanding of the consumer. Why would we go back to pretending retailing is best measured using old metrics?

Casey Golden
BrainTrust
16 days 23 hours ago

Agree. No longer are demographics, psychographics, preferences, influences, etc. considered only “inferred data.”

Mark Ryski
BrainTrust

Yes, this is a good step, but it’s only one step. Transactional data only accounts for the shoppers who made a purchase — what about the shoppers who visited and left without buying? Store traffic data is one of the most — if not the most — reliable activity signals of store level activity. Furthermore, store traffic quantifies a store’s sales opportunity. In addition to traffic, shopper conversion rates provide important insights into how effectively the store is doing at turning store visits into sales. Ultimately, retailers need to combine transactional data with other, non-transaction data to find the insight gold.

Verlin Youd
BrainTrust

Absolutely! Traffic and conversion are a good start to understanding all of the behavior that goes on prior to checkout. New solutions are finally providing a useful window into that behavior.

Matthew Pavich
BrainTrust

Understanding the buying behavior of your customers is definitely foundational to running a successful business. Not only is it invaluable in understanding how they have behaved in the past, but it is the most critical input into predicting what they will do in the future. Every strategy should be heavily informed by customer data with the right sophistication and AI being used to identify key shifts and trends in how they behave coupled with the technology and processes to evolve accordingly.

Dion Kenney
BrainTrust
17 days 1 hour ago

Yes, the customer-base audit of historic purchases is the baseline, organizing the results into clusters of meaningful demographics and psychographics. And it should be considered mandatory today, with the easy access to both the Big Data and data analytics tools necessary to perform it. But beyond the baseline, the data-driven marketer should also include a correlation analysis of “when purchased vs exogenous phenomenon” to identify meaningful variables and better predict future purchases against them.

Dave Wendland
BrainTrust

Transactional data is certainly a big part of the equation. However this needs to be combined with qualitative data that examines behaviors not just transactions.

Verlin Youd
BrainTrust

Yes, as well as the quantitative data that examines and highlights all of the behavior prior to transaction!

Nikki Baird
BrainTrust

Well, you have to first take into account whether you actually know your customers or not and can build a basket history over time. Retail’s ability to do that has grown significantly in this area over the last 20 years, but also took some hits (Apple tracking, for example, and the reduced ability to ID shoppers if they have not volunteered that identification). But otherwise, definitely – how can you claim to be customer-centric if you don’t know what your customers are actually buying? The only drawback to looking at purchase data is if it takes up so much of your time that you don’t look at other things. Purchase history is the center, but you need to look at leading indicators (social sentiment, for example) and lagging indicators too (customer sat), to see how it all comes together.

Lee Peterson
BrainTrust

A former CEO of mine used to say, “you don’t know anything until that register rings,” which is true. But aside from that, I really believe in observational research. Whether it comes from store managers, buyers, execs of all stripes or the CEOs themselves, nothing beats actual on-site experience. And that goes for online too. How many times have you had an experience with a retailer and said to yourself, “wonder if the CEO knows how hard it is to do X?” In short, if they don’t “live” with their own brand, they should.

Mark Self
BrainTrust

I guess I need to read the book. Haven’t retailers like Walmart analyzed product movement, basket sizes and content by leveraging analytics, Big Data, transaction logs, and the web, for some time now? As an aside, when there are available tools and a framework available for this assessment, that usually screams “consulting engagement” to me.

Buying behavior over time does change — people age and their shopping list changes.

Coming back to the question — the benefits of leveraging purchasing data can be great, IF the retailer makes the commitment necessary in talent and tools to attain these “real time views.” Many times that commitment and investment is “under sized” and the benefits are not immediate, making a strategy change and investment like the book suggests difficult, especially for publicly traded companies.

Peter Charness
BrainTrust

That’s so last century. OK, it depends on the type of product and transactional data is clearly critical. But the customer journey starts well before the “buy” button, and doesn’t end there. If all you have is transaction logs well then have at it, but today understanding customer behavior demands much, much more.

David Spear
BrainTrust

Sure, every company should understand their customers and the customer-base audit, as described by Prof. Fader, is a solid first move, but there are many more layers that need to be peeled back. Descriptive analytics only tells you what happened (backward looking), but the real treasure is when insights are generated with predictive (what you could do) and prescriptive (what you should do) analytics. This means massive data sets sourced internally and externally.

Ananda Chakravarty
BrainTrust

I’m not familiar with the details of Prof. Fader’s focus on transactional data — but almost all retailers capture and review t-logs and other transactional data. The challenge is that it provides limited information about the customer when taken in isolation. Sales, units, products and timestamps are it. But this is common information and every retailer uses it. Merchandising for a customer-centric view needs more than this to offer a competitive advantage. More importantly, if retailers use only this type of information – they will only have the same type of results. This is a recipe for zero change and zero growth. Lastly, the data offered is all historic data. They will never be able to assess future scenarios based on past data alone. This mirrors an adage from Wall Street about past performance. Retailers need to strike a balance between transactional and other opportunities to collect customer data such as loyalty programs, CSAT surveys and NPS scores. Customer-centricity is much more than transactional data.

James Tenser
BrainTrust

There’s plenty to like about making more systematic use of transaction data to understand retail business health. The customer-base audit is a natural adjunct to the widely-used Lifetime Customer Value discipline.

I have mad respect for Dr. Fader and I like this “5 lenses” cohort analysis approach that seeks to establish and track some fundamental facts about purchase trends. Are newer customers buying differently compared with older customers? Are purchases of certain products associated with greater LTV? How is your customer acquisition trending? Are existing customers changing their purchase behavior?

These descriptive measures belong on the BI dashboards of senior management. They will help bring key trends to the surface. Getting to “why” is the role of other marketing methods, including attitudinal, sentiment, loyalty and behaviors outside the retailer’s “walled garden.”

DeAnn Campbell
BrainTrust

Transactional data tells you what customers ultimately buy, but that’s only a small part of the story. Are they buying it because their preferred product is out of stock – meaning they likely won’t buy it again? Are they spending a lot of time considering a more expensive product but ultimately defer to price and leave feeling naggingly unsatisfied? Are shoppers avoiding a part of the store because of its location or look, but would buy the products if they were better positioned? This is way more important data to gather because ultimately it drives customer satisfaction, repeat business and long term success.

Brad Halverson
Guest

Spot on. Getting an understanding of less frequent shoppers, or even shoppers who haven’t adopted your business is just as important. From here, you’ll be able to understand any barriers and ultimate solutions for transformation.

Ryan Mathews
BrainTrust

Like Dave Bruno I would need to read the actual study, but my first reaction is that in attempting to correct one flaw the authors have defaulted to another, possibly worse one. Hardie’s assertion that, “We’re not interested in who these customers are, not what they think, or their attitude,” is nothing short of — well — given The RetailWire Golden Rule, I’ll just say massively shortsighted. Here’s the logic flaw. Transactional data always is a measurement of a forced choice, the option the consumer selected from among the objects for sale. If I’m a Count Chocula fanatic but the supermarket is out-of-stock I may grab a box of Lucky Charms. A purely transactional analysis will assume I like Lucky Charms. Like Dave, without reading the whole study, I think this appears to be the marketing equivalent of “mutton dressed as lamb” — something old wrapped in the veneer of something new.

Casey Golden
BrainTrust
16 days 23 hours ago

Can you really be customer-centric if you build a strategy only on transactions and not accounting for customer decision making process and why they choose this over that? I’ve never found long-term success in a complete top-down strategy. This quote really hits a chord: “We’re not interested in who these customers are, not what they think, or their attitude.”