FEATURE5 August 2011

This month we... browsed a virtual supermarket

Robert Bain explores a simulated supermarket used to research products and store designs.

Trips to the supermarket don’t tend to be very memorable. Retail marketers spend their time coming up with ways to appeal to shoppers who aren’t thinking very hard about their choices – at least not consciously.

But this trip is a bit different. I’m standing outside a branch of Grocery Inc in Somewheresville, USA. It doesn’t sound much like a real supermarket because it isn’t: I’m actually sitting on my sofa with my laptop, and the store before me is a virtual one.

Keeping it real
This digital supermarket was built by InContext Solutions, a two-year old company that specialises in shopping simulations for market research.

For manufacturers and retailers, knowing how people are likely to react to a product, display, promotion or store layout before investing in the change is hugely valuable. But live tests require time, money and risk. Simulations offer a way around this – in a virtual environment you can fail as much as you want, and you’ll only use virtual dollars.

“In addition to the sales data we can gather behavioural data like how long it took them to find a particular item or where they went as they were looking for things”

Rich Scamehorn, InContext Solutions

Rich Scamehorn, one of the founders of InContext, knows this from his previous work on in-store research at food manufacturer General Mills. Four or five years ago, he says, there were two types of virtual store research available. Companies such as Red Dot Square (which was later bought by Kantar) offered high-tech 3D simulations, but interviews had to be conducted in person, making it a pricey option. Those systems that could run over the internet tended to be “very 2D in nature”, says Scamehorn.

As the power of respondents’ home computers improved, it became possible to deliver high-quality 3D simulations over the web, and Scamehorn saw a gap in the market that the existing providers weren’t filling. He teamed up with Bob Gillespie and Tracey Wiedmeyer, whose backgrounds were in software and tech consulting, to launch InContext. They now have 35 staff, of whom 14 are dedicated to developing the technology and making sure it works. The company says tests using its virtual stores can be done in about a third of the time and a quarter of the cost of real store tests.

Just browsing
To arrive at the entrance of Grocery Inc, which is based on a standard Walmart layout, I have followed a link in an online survey, installed a browser plug-in to stream the images and waited a few minutes for the simulation itself to load.

As soon as I click Start, I am whisked through the glass doors and over to an aisle near the back right corner of the store. This particular simulation is designed to research oral care products, so I’ve only got the run of one aisle, but there’s plenty to play with. I walk around using the arrow keys, tapping the spacebar to look up and down and clicking the mouse to pick things up. It’s a lot like a first-person shooter game – only with fewer monsters and more consumer products. Imagine a supermarket in Doom before the hell demons arrived.

Like a lot of computer simulations, it all looks a bit too bright, clean and tidy to be real. But the level of detail is impressive.

Dozens of products fill the shelves, with each package clearly identifiable. The signage, the ceiling lights, the shelf fixtures, the floor tiles, the cash registers – everything is there.

I’m attracted first of all to the exotic US brands that I don’t recognise from home, including a mouthwash called Scope and a toothpaste called Topol (not to be confused with the actor who played a crazed scientist in Flash Gordon). That’s my post-rationalisation, anyway. InContext will no doubt have other theories about factors that influenced what I did – the height of the shelves, the package designs, the position of the products in the aisle.

When I click on a tube of Topol, a close-up fills the screen. I can rotate it and zoom in and out. It’s an accurate 3D model of the real product, detailed enough that I can clearly read the ingredients, disclaimers and manufacturer’s address. Photographing the packaging and building 3D models of all these items is one of the most time-consuming parts of the process, Scamehorn explains.

Topol, I learn, is “now 30% larger”, with a net weight of 3.5 oz. Tempting, but it seems to be targeted at people who either smoke heavily or drink large quantities of black coffee, so I put it back and have a look at what else is on offer.

What will InContext be able to glean from my behaviour? If Topol were launching a new product they might run simulations with several different package designs, comparing the resulting sales and studying whether their new offering ate into competitors’ sales or cannibalised those of its sister brands.

“In addition to the sales data we can gather behavioural data like how long it took them to find a particular item or where they went as they were looking for things,” says Scamehorn. “Did they pick up certain items, did they put them back? What did they purchase? And after they’re done we have a follow-up survey so we can get some attitudinal measures about why they did what they did.”

“One of the beautiful parts of virtual store research is that you can test a lot of different scenarios without the cost of generating physical versions”

Rich Scamehorn, InContext Solutions

Retail realities
Companies like InContext find themselves in the middle of a relationship between store owners and product makers. Historically, Scamehorn says, it has been the manufacturers who have driven this sort of research. “A lot of manufacturers are trying to get something to allow them to sell into the retailer,” he explains. But he believes things are changing. As more retailers launch their own ‘private label’ ranges and expand their marketing activities, they’re getting “more savvy” about retail research.

The web-based approach is not just cheaper, it is also far more flexible, says Scamehorn. “It really opens up the ability to have larger sample sizes and analyse the data in more ways. One of the beautiful parts of virtual store research is that you can test a lot of different scenarios without the cost of generating physical versions of a lot of different packages or aisle fixtures or whatever it might be. That is a huge deal. From the manufacturers’ side it’s also more secretive, because nobody can see what you’re doing.”

So far, InContext’s tool has been used by clients in the US, Chile, Spain and Italy. And there’s no reason for it to be limited to recreating retail environments.

“We’re exploring some things around sporting venues, looking at helping them show people what advertising opportunities are available,” says Scamehorn. “Currently they would have to bring someone to the venue – with our tool they can have that meeting either over the internet, or at any location in the world, and show them what it looks like in our virtual arena, walking through and seeing how it looks from different angles. It would be great for the Premier League.”

1 Comment

13 years ago

How 'REAL' is this? Are respondents shoppers or browsers? What does the deodorant/shampoo smell like? Can you compare products 2 at a time for ingredients? If you watch shoppers in-store for 10mins, you will see how many THINGS shoppers do pick up, put back, shake, smell, check list, show to partner etc before buying. Cheap YES, realistic mmmhh. Maybe a staging post before real trial but I wouldn't risk my research job on the insights and recommendations.

Like Report