If you had a perfect measurement system for in-store behavior, what would it do? It would track every shopper. It would track every Associate. The tracking would be highly detailed – down to say the 1 second level in granularity – and it would be positionally accurate – placing each person at their exact location in the store. It would also be comprehensive – following their entire journey.

That’s the core of an ideal measurement system for people-tracking and geolocation analytics.

But there are some nice to haves too. It would be nice to know which Associate is which. It would be nice to know if a shopper is a regular customer. It would be nice to know what  a shopper looked at, not just where they were (eye-tracking) and it would cool to know whether they touched or tried a specific product. And while we’re at it, it would be nice to something about the shoppers. Their gender, their age, their group size, stuff like that.

And, of course, the best solution would be inexpensive to buy and easy to deploy.

In this recent series of posts, we’ve been looking at the measurement capabilities of camera and electronics for people-tracking. If you rate each versus this Platonic ideal of a measurement system, it would look like this:

There are a couple of pretty obvious takeaways from this list. First, if you just look at the core requirements – population, constant tracking and positional accuracy – camera crushes electronics. It’s better in every respect. But that isn’t all you should take-away. Note how opposite the list is. There’s almost no respect in which the two types of tracking perform similarly. What camera is good at, electronics isn’t. And where electronic excels (full journey, repeat visits, Associate identification), camera flounders.

It’s interesting that the two technologies are almost perfectly complementary. Camera is great for full population. Electronics sucks for that. Camera is highly positional accurate so you can do detailed queue analytics and merchandising tests. Electronics isn’t and can’t support those applications. Camera gives you a lot of nice to haves like gender and height and grouping and potentially additional detailed behavioral data. Electronics doesn’t do any of that. On the other hand, even with full coverage, camera will have blind spots and will regularly lose track of shoppers. So it can’t seamlessly track the full journey. Electronics can. And putting camera up to cover large interstitial elements like stairs and walkways is expensive and sometimes impossible. Electronics works well in those situations. Camera can’t track repeat visitors either. But electronics can. And while both camera and electronics can potentially track Associates, similar dynamics exist. Camera can do very detailed exact process analysis. But it can’t track specific individuals. Electronics can’t track detailed paths and process execution, but it can identify specific individuals.

So it looks like no matter which technology you pick, you’re going to have gaps. With electronics you’re going to have big gaps including population tracking and any kind of fine-grained detailed. But even with full camera coverage, you won’t get all the information about the journey that you’d really like.

When we started Digital Mortar, we used electronics almost exclusively. But the huge gaps in the data drove us crazy. Gradually, we started to deploy more and more camera sensors. But when we went to full camera coverage sans electronics, we found that the gaps in full journey tracking were less than ideal.

Which is why in our most recent (and best) implementations, we do extensive camera coverage AND we deploy a significant number of electronic sensors. Camera covers all the main entries and all the main merchandising areas. We use a lot of camera. But we use electronics to cover overall flow including all the interstitials and repeat visitor tracking and for Associate identification. We even blend the two streams so that, for example, an Associate with a blue-tooth pinger is automatically identified when they come through a door and then the camera tracks them on the floor.

This integration of electronic and camera sensor technology is complex and has its challenges. But only by putting the two together have we been able to create a measurement system that answers the full-range of people-tracking questions at the enterprise level. Camera does most of the heavy measurement lifting, but electronics fills in the gaps and provides the big picture flow measurement.

Electronics and Camera are the “odd-couple” of people-tracking technologies. They don’t do anything alike, but they work surprisingly well together.