In-store success is often won at the shelf. For a shopper to choose your brand out of a sea of competitors, your products need to make it to the shelf, be displayed correctly, be re-stocked when needed, and many other crucial steps.
That’s why it’s a no-brainer to focus your data collection efforts on the store shelf. How can you win the sale if your products aren’t set up properly? If they’re out-of-stock, or the display is broken, or your competitors are taking over more share of the shelf? High-quality in-store data can alert you to problems and help you prioritize your fixes.
However, the standard approach is a manual one. You’ll send in some merchandisers or field reps. They’ll hit their target stores, manually collect some shelf-level data, and start building a plan from that. This approach is prone to errors. Plus, it’s time-consuming and not scalable. It’s nearly impossible to have enough people to visit every store, and odds are they’re in good stores along with the bad.
Today, there’s an automated solution that addresses problems of scalable, cost-effective in-store data: image recognition. What is image recognition? Let’s find out.
Image Recognition Defined
Image recognition is a technological innovation, for sure, but it’s also added value for your business. Image recognition is an algorithm that can process an image or video and decode the contents of that visual media. The tool can identify what is in the image—whether that’s a specific product, a brand, a category, or something else.
Consider this example: You’re a cereal brand and need to know whether your different cereals are displayed correctly on the shelf. You have a field rep go into the store and take a picture. It’s quick—in and out. Then, that picture is fed through image recognition. The software determines what was actually photographed—it identifies the specific products, the packaging, the sizes, the competitor cereals, and more. It packages that data up into a nice visualization and you can quickly tell what’s going on at this shelf.
A big win from this example is for your field rep. They just took a picture, they didn’t have to stop and analyze the shelf, go through a manual checklist, and so on. They were busy doing other value-driving activities at the store instead.
Why Is Image Recognition Important?
Time savings are a big reason why image recognition is important in retail. It puts the onus on the tool to decipher what’s going on in an image, not on your team. It’s a lot faster to take a few pictures than it is to answer a survey or catalog every detail of the shelf.
In addition, image recognition is also a boost to data quality. You can combine insights from on-the-ground reps and other sources with image recognition applications to build a better picture of the store. Reps can capture other data points, for example. Or, you can have the AI and machine learning scan the images for other attributes or build a completely different picture, like stitch multiple snapshots into a panorama.
Furthermore, image recognition is also an important part of the process for Wiser Solutions and how we deliver value to our customers. We want to support your field teams as they capture in-store data, including pictures. Using image recognition, we can shorten the amount of time needed to gather actionable intelligence. Beyond that, we can supplement your data with other insights so your reps can have smarter conversations with stores.
All in all, this technology can contribute in a variety of ways to your ultimate end goal—owning and controlling more space inside the store. Plus, you get the possibility of less expensive data collection that is higher quality than the manual alternative. That’s a win-win!
Image recognition is an algorithm that can process an image or video and decode the contents of that visual media.
Use Cases for Image Recognition In Retail
How can you use this technology to provide value for your business? Here are few examples of some use cases in the retail space:
- Shelf health—Is your brand’s in-store presence healthy? Products need to be on the shelf, pricing needs to be right, adjacencies should be as expected, and more. Image recognition can identify a lot of these data points inside pictures of store shelves.
- Display compliance—Are your in-store displays set up properly? Are they broken? You’ve spent a lot of money planning and building that display, so it better be functioning as intended on the ground. Image recognition can look for specific signs of a healthy or unhealthy display.
- Competitive insights—What are competitors doing in the store? You need to know what they’re up to, whether it’s their location on the shelf or new products, displays, or pricing. Image recognition can identify competitors in pictures of the store.
- Planogram compliance—Are your planograms executed as planned? Like with displays, you spent a lot of time and effort building those planograms. Image recognition can compare on-the-ground photos to a sample planogram to determine whether or not they match.
What an Image Recognition Algorithm Needs to Succeed
At its core, image recognition software is all about more—more data, more quality, more scale. However, it’s not a magic solution and does come with some considerations.
For starters, the artificial intelligence is not a magic solution. It’s only as good as its inputs. You need to start with high-quality imagery for it to work. This is a big focus at Wiser, as we work with our mystery shoppers and your field teams on how to take the best quality pictures of the store. If you’re using your own teams, for instance, make sure they have the tools to take good images and the training to know what to avoid (bad lighting, crowded aisles, steep angles of the shelf).
Lastly, you need a solution that ends with actionable insights. It’s great to be able to pinpoint what’s going on in a photo, but you have to be able to do something with that information. Look for a platform that has great business intelligence and data visualizations. Make sure the outputs are tailored to your needs. It also has to be completed in a relatively short amount of time—the point is to streamline in-store data collection, so it’s not that great if image recognition takes longer than your manual workflows.