I was invited to speak at Mobile Marketing Magazine’s mRetail Summit last week to talk about the history of location-based services. Some of you may or may not know that I spent my early career in fashion retail and then some years later in 2000, joined the mobile scene at ZagMe, where we sent text messages to shoppers at Lakeside and Bluewater shopping malls while people were actually out shopping. My retail experience certainly came in very handy there. Anyway, I was lucky enough to be, not only at the birth of mobile marketing, but also at the birth of mobile location-based services.
I’m not sure my slides will mean much without the dialogue, but they’re here if you want to take a look.
I suppose the key learnings for me from that time are
- Customers wanted a permission-based offers channel for their local shopping mall
- Once they’d signed up to the service, people assumed we knew where they were (we didn’t, we spoofed the location bit) and expected us to send them relevant messages
- We were at least 10 years ahead of our time
- It’s not about the technology, it’s about the application of technology. Understanding how to communicate with customers is essential. Understanding how retailers and retail staff operate is critical.
- The ZagMe concept is still sound, but despite several services attempting to do something similar in the last 10 years, they’ve all failed. My guess is that it’s because they focussed on the technology and thought that retailers would change the way they do things to adapt to the technology rather than the other way around.
- Location information is mostly useless without some context
- Past behaviour is not a good predictor of future behaviour. A recent example to demonstrate this… I’m going to Barcelona next month. I searched extensively on Google to find my accommodation. Now the only ads I see are adverts for accommodation in Barcelona. If you really did understand my behaviour, you’d know that the time to serve me those ads is in December and early January. By the end of January it’s too late. But behaviour predictors don’t have that level of granularity. But I guess that conversation is for another day.