FEATURE1 January 2012

The path to purchase


Technology is helping bricks-and-mortar retailers get better at understanding shopper behaviour – just like their digital counterparts. Tim Phillips enters the world of advanced shopper research.


These are not happy times for retailers. In the UK around 20 shops close every day and one in seven shops in town centres is vacant. So for retailers the need to know what shoppers are thinking – and more importantly what they are doing – has never been more pressing.

Yet shopper research has always been characterised by flaws: a lack of ability to tie macro spending trends to individual shopping decisions; difficulty knowing what customers do when they are not in the store; no firm knowledge of the effect or potential of promotions especially when predicting long-term behaviour. Research is also constrained by the difference between how shoppers recall and interpret their behaviour and the behaviour itself.

Loyalty card data has opened up the field of analysis based on revealed rather than stated customer preference. It is a rich source of insight, but is limited by how widely people use the cards themselves. On the other hand, new advances in shopper research are helping to track the individual behaviour of shoppers, creating a better picture of the customer’s decision-making process for retailers.

Perhaps the most important tool in all of this has been the mobile phone – and often the customers don’t even know that it’s the phone that makes the difference. So it is with Path Intelligence, where CEO Sharon Biggar has worked with UK retail brands to track their customers, simply by passively observing where mobile phones (which move with them) are at any time and matching that aggregated tracking data to departmental and store sales. An example: TK Maxx could not understand why one of its stores had a conversion rate of only 18% when the average was 23%. Path Intelligence found that one in five customers simply used the store as a shortcut from one part of the mall to another. Disregarding those customers showed the store performing in line with expectations. But also, given that they were entering the store, experimenting with reconfigured layout successfully created a new stream of revenue by converting passers-by into buyers.

Amazonian principles
This shows two aspects of Path Intelligence’s model that borrow from the online world: the desire to track customers through their entire shopping journey, not just analyse sales, and a willingness to conduct controlled experiments to improve the results in a collaborative process of continuous improvement.

Biggar says: “Online retailers have a wealth of information about how shoppers behave. What we have seen in the UK in the four years to 2011 is that online shopping has grown by 215%, the high street only 4%. One reason is that, offline, we simply don’t have the data we need.”

Path’s data can be turned into heat maps, traffic flow diagrams or affinity maps, but what matters for Biggar is that they become the starting point for an investigation, not the result of it. “Amazon and eBay, for example, have experimentation teams, and some of our clients are doing the same thing in store. They will change layout on a weekly basis to see what works,” she says. Path Intelligence, which prefers to do the data analysis rather than dump and run, nevertheless does not ask customers what they want to do. Instead it simply observes their actions and creates models that link behaviour to revenue, without an attitudinal interpretation. “Our view is that market research isn’t reliable if you ask someone to remember something,” Biggar says. One of their key results from a long-term survey: over nine months, a 1% increase in dwell time is linked to a 1.3% rise in sales.

“Following the shopper on a safari means giving the client a fresh pair of eyes through which to see the shopping journey”

Technical limitations
Investing in technology is essential, but only if it answers a real question. Criticisms of shopper research techniques such as neuroscience and eye tracking are not just that results can be unreliable – all MR has error bars – but that a result does not necessarily suggest a clear course of action. That’s why Marketing Sciences likes to use advanced techniques like eye tracking to enhance its existing methods, but not use them on their own.

“Clients don’t care about methodology; they care about insight,” says Ian Ralph, a director whose background in the tech sector makes him aware of the limitations of technology as well as its potential.

Associate director Rachel Rockey explains that ethnographic techniques like “shopper safaris” (their term) can use mobile phone tracking plus eye tracking to help enhance a shopper log, not replace it.

“We will conduct quant research to investigate awareness and we follow it up in-store using eye tracking, so that when a respondent said they saw something, we can see whether they really did notice it in-store,” she says. “The retail environment today is an overload for consumers. Christmas exemplifies that – and we know that customers tend to filter out a lot of the messages. Do they ever look up? We find, no. And they don’t look at the floor either.”

The results help to start discussions with clients. Following the shopper on a safari means giving the client a fresh pair of eyes through which to see the shopping journey – and sometimes to conclude: “Of course the customer wouldn’t notice that.”

The start of the journey
Other agencies use the mobile phone to create reliable customer stories, now starting from before they reach the store. Synovate is one of the agencies using mobile ethnography for shopper research, which starts with passive monitoring of the pre-sales research that respondents do (it also monitors use of the home computer).

“It uses the phone as a sort of passive meter of what the customer does: it is watching you, so you have to be very transparent with the respondent about what you are doing. But using the computer in the same way, we are able to track all the research that shoppers do before they buy a product,” says Thomas Edwards, global head of mobile solutions.

Tracking the use of search engines, blogs and price comparison sites provides a rich set of data about how careful shoppers plan their trip. Then software developers are needed to create applications to integrate biometric data from ‘smart plasters’ which monitor the vital signs of respondents as they go about their shopping trips, with the data automatically transmitted via Bluetooth to a mobile.

“It is more of a client collaboration when you are doing things that have never been done before. If you send a client a specification it’s difficult to get them to read it. We have to sit down with them and make sure they know what we are doing – because it switches the focus from market research to software development,” Edwards says.

Clients need hand-holding, not least because the data doesn’t support a simple interpretation. “The shopper typically goes back and forth from one site to another, changing his or her mind all the time. To be honest, it’s often a mess,” Edwards admits, “A person will go to 25 websites. It’s hard to show those charts to clients.”

Edwards wants retailers to commit to tracking as many of their customers as possible and using as much of the data as possible. He wants to get away from averages and panels, he says.

Watching closely
Another shopper research specialist that wants to use all of the data all of the time is ShopperTrak. As one of the largest and longest-established shopper research specialists, ShopperTrak has been videoing shoppers since the technology to do so was in its infancy. Now, says Todd Starcevich, chief client operations officer, software algorithms make it possible for the cameras to continuously monitor the customer around the store.

“What ShopperTrak can answer is the question of how you turn as many visitors into buyers as possible,” he says. “We have forced retailers to get smart about how they build their experiences.”

By building a 3D model of the shopper’s journey, using every journey from every shopper, ShopperTrak has a much clearer idea of changing shopper patterns of behaviour. One development in behaviour that Starcevich observes is the rise in importance of the “surgical shopper”. Also noted by Synovate, the surgical shopper doesn’t shop without research. By the time he or she has reached the shop, they have made their choice.

“Their spending decision is already made,” he says. “The dynamics are different. You need to know that when they show up at the door of the shop you have the product they need at the time they need it. We look at every person in the shop, and ask the retailer: did you have the number of people you were expecting? Did you have the stock they need? Capturing everybody is a key foundational component so you can rely on the sample set, rely on the confidence intervals and say that this is really what’s happening to the total population – because if you are sampling only four or five stores you will get inconsistencies.”

So shopper research is expanding into areas not covered by traditional product research: are the retailers matching stock and staffing to demand? The problem to solve isn’t so much how to divert the shopper’s attention as how to make sure that there are minimum obstacles to putting the surgical shopper in touch with the product that has already been selected. This means revisions of store layout, stock and staff allocation.

“There’s a huge difference between being an anonymous respondent and an active participant who willingly trades information for a better shopping experience”

Foursquare for research
Increased knowledge of what the shopper does, even when not buying, may one day mean that anonymity is a quaint historical idea. Yet for all these applications, there’s a huge difference between being an anonymous respondent, even one who is passively tracked, and an active participant who willingly trades information (about past purchases, or even the fact that they have just entered the store) for a better shopping experience.

Surgical shoppers have an incentive to share information about their behaviour – they, after all, just want the shortest route to the best product. A frontier in shopper research for Edwards and Synovate will be the opportunity to use all the data available on every customer to help them directly.

“If we can do what FourSquare does in a way that doesn’t annoy the consumer, you will get data that’s better than any other type. Once you have it figured out it could be very exciting for retailers.”

Ultimately, much shopper research can blur the line between research and marketing. By making it possible to research individual shoppers, and to act on that research, then anonymity may no longer be an essential – or even desirable – feature of research.

Advanced shopper research offers passive monitoring, an escape from the dead end of hearing shoppers post-rationalise their actions, experiments and real-time feedback, and a tantalising glimpse into a future where the interaction with a customer may be practical to do one to one. But, to be successful, it is also focused on affordable technology that answers real questions rather than gimmicks.

“We know we can provide data that can be predictive,” says Starcevich at ShopperTrak. “I just hope we never get classified as a weird gizmo company.”

Low-tech, leading edge

Leading-edge shopper research uses behavioural cues to find deeper motivations for buying decisions than those the shopper may be aware of. This can be done by lighting up parts of the shopper’s brain, but a simpler way is typified by Wendy Lanchin, planning and strategy director at The Marketing Store (TMS).

“Observational research will tell us how shoppers engage with the in-store environment: how they move around the store and how they engage with fixtures and display,” she says. “It can only show us the what of shopper behaviour and not the why. Exit interviews and accompanied shops are often used to tease out more insight into observed behaviour, but these will often elicit little more than the obvious.”

Lanchin wants to capture an emotional response, so she is working with BrainJuicer, which has built a model of emotional response using its FaceTrace methodology. This maps consumer emotions to seven recognisable states, including sadness, fear and contempt (if your shoppers display all three, you have a problem).

BrainJuicer uses these responses to evaluate advertising; TMS uses them to evaluate shopping experiences by tapping into automatic emotions that reveal their state of mind. In this way, shoppers can report their state more effectively using low-tech tools.

TMS is reporting on the first stages of its research at the Esomar Shopper Marketing conference in February. Lanchin is confident that the results show shoppers go on “clear, well-defined emotional journeys during the course of their shopping trips, influenced by measurable and actionable triggers.”

And you thought you were just looking for the deli counter.


12 years ago

Great marketing and eye-catching merchandising will always result in sales. Granted that today, a shopper may see a product in a retail store and purchase it online - BUT the marketing must be done at retail where people can see and touch the product before buying. The issue is how to support the retail "showroom" and price the product so the sale is immediate, not later online.

Like Report

12 years ago

This is a really good article. Eric, like your closing sentence....."The issue is how to support the retail "showroom" and price the product so the sale is immediate, not later online".

Like Report