What we learned from 10+ years in location intelligence

Kjartan Slette · Co-founder, Pine59

More than ten years ago, Thomas and I started working with location data.

A lot has happened since then.

We've built products that worked, and products that didn't. We've worked with some of the world's largest companies and with startups trying something for the first time. We've watched technologies appear and disappear (iBeacon anyone?), privacy expectations change, regulations arrive, and the location data industry reinvent itself several times.

We also went from building Unacast to eventually starting again with Pine59.

So when we started Pine59, we weren't starting from scratch. We were starting with more than a decade of things we'd learned, including quite a few things we learned the hard way.

Here are some of them.

1. Most customers don't actually want the location data itself

They want answers:

• Where should we open our next store?
• Why is this location outperforming that one?
• Where are our visitors coming from?
• How is this area changing?
• Where should infrastructure be built?

Raw location data can help answer those questions, but the customer rarely wakes up wanting another enormous dataset to store, process, govern and understand.

This realization has shaped a lot of how we think about Pine59.

Our job isn't to move as much data as possible from our infrastructure into yours. It's to do the difficult work upstream and turn location data into useful intelligence about places and populations.

That's also why we increasingly use the term Aggregated Location Intelligence.

The intelligence itself is the product. The underlying observations are simply how you get there.

2. Privacy isn't something you add later

The location data industry has changed enormously since we started.

So has the regulatory environment.

GDPR arrived in Europe. Privacy legislation is developing across the US. Platforms have changed how location permissions work. Consumers, companies and regulators have become much more aware of how data about people's movements can be used.

One lesson from watching all of this happen is that privacy architecture matters much more than privacy language.

If your entire product depends on receiving and processing individual device-level location data, changing that architecture later can be extremely difficult.

That's one of the reasons Pine59 is built around pre-aggregated data. The data is aggregated before it reaches us, and neither Pine59 nor our customers need to receive the underlying device-level location data to use our products.

We think that's a much better foundation for the next ten years than trying to predict exactly what the next regulation will say.

3. The dashboard is often just the beginning

We love a good dashboard. We built one.

For many customers, being able to explore a location, compare places and understand a market visually is exactly what they need.

But we've also learned that some of the most valuable uses of location intelligence happen when nobody is looking at a location intelligence dashboard at all.

The data is inside a retailer's operating system. A real estate model. A consultancy's methodology. An AI product. A planning tool. An application we've never heard of before.

At that point, location intelligence has stopped being something you log into and started becoming infrastructure.

That's why APIs and flexible data access have become increasingly important to how we think about Pine59.

We want customers to use our products directly. But we also want them to build products we will never see.

4. Customers will always come up with better use cases than you do

This might be my favorite lesson.

You can spend months building a product, creating use cases and carefully explaining what people should do with it.

Then a customer walks into a meeting and says: “Could we use it for this?”

And suddenly everyone in the room is thinking about something completely different.

We've recently seen people explore location intelligence for everything from emergency response and national statistics to robots, global retail operations and combining outside-store movement with what happens inside the store.

Some of those ideas will become products. Some won't. Others will probably turn into something completely different.

That's exactly how innovation is supposed to work.

Our job isn't to predict every possible application of Pine59.

It's to build a reliable foundation and make it easy enough for other people to experiment.

5. More data isn't necessarily always better data

This took us a while to fully appreciate.

In location intelligence, big numbers are seductive. Billions of signals. Millions of devices. Massive datasets.

But the size of a dataset tells you surprisingly little about whether it can answer a question reliably.

What matters is whether the data is representative, consistent over time, geographically useful and processed with a methodology you understand.

Ten billion observations from the wrong population don't magically become good statistics because there are ten billion of them.

Today, we're much more interested in the quality and consistency of the signal and panels than the size of the number in the sales presentation.

And perhaps the biggest lesson: keep learning

After more than a decade in location intelligence, it's tempting to think we know the industry pretty well.

We do.

But one of the reasons starting Pine59 has been so much fun is realizing how much is changing again.

AI is changing how software gets built and how data gets consumed. Privacy regulation continues to evolve. New sources of data are emerging. Products that would have required a large engineering team a few years ago can now be prototyped in days.

So perhaps the most useful thing we've learned in ten years is not to get too attached to what worked during the previous ten.

Keep the experience. Question the assumptions.

And keep building!