5 lessons learned from one of Australia’s largest beacon deployments

James Irving /

We recently completed a large deployment of over 300 beacons in a major shopping centre in Melbourne. Whilst we can’t talk about the specifics just yet (stay tuned), we can share some of our key learnings with you. The scale and complexity of the centre layout offered some new challenges to overcome. So here we go, 5 things we learned from the deployment:

1. Plan and program your beacons before the deployment

For one of our first ever beacon deployments we had planned to program our beacons prior to heading to the site but circumstances prevented us doing so and we had to program each beacon on site. I can tell you now that beacon management apps don’t like 300 beacons in range whilst trying to program one of them. Trying to work in this way is extremely slow and tedious and I don’t recommend it for anyone.

Tip: Keep your beacons in a microwave to prevent them from broadcasting signals while you test and program other beacons.

IMG_1894.JPG

Lighthouse team doing some deployment planning

2. Test beacon ranges in the live environment before deciding on the range and power settings

The beacons we used could be programmed in increments of -40dbm (~3m), -20dbm (~12m), and -16dbm (~18m). During testing in our office building we found the approximation of the range fairly accurate and consistent. Based on this testing we decided on a -20dbm (~12m) range. However, when we installed the beacons we found that in a shopping centre full of reflective surfaces such as tiles and glass the range became closer to 50m. This resulted in us having to reprogram all of our beacons to get the desired range. Likewise, carpet or crowded spaces will significantly reduce the range of the beacon.

IMG_1884.JPG

Beacons – programmed and ready for deployment

3. Two beacons are better than one (sometimes).

A wide hallway or room has better coverage and more consistent events when we used two beacons at a lower range rather than one beacon at a high range attempting to cover the entire area. We found out that humans are often effective radio frequency shields and a single beacon in the centre of a wide walkway would often not be picked up if the user had their device in their pocket. Having one beacon on each side of the walkway was much more reliable solution.

4. Beacons at medium to far range (~12m+) cause issues for multi level buildings

If you are deploying within a multi level building then there’s a high chance you won’t want a user’s device picking up beacons from the level above or below. The signal from beacons programmed at anything at or above about -20dbm will have little trouble passing straight through solid concrete floors and being detected by devices on the next floor. To address this issue we used several beacons programmed at a lower range to create the zone we wanted without overlapping into the floor above.

5. A Beaconater 2000™ is far better than a ladder.

This is our trade secret we’re giving away here so be grateful for that.

In our first deployments we carried a ladder with us so we could reach high points in the centre. This was an extremely slow process and called for a unique solution. We considered using a scissor lift but that too would have been a slow process. Enter the Beaconater 2000™, created by none other than Inlight GM, Patrick Carne. It’s a pool cleaner crossed with a grasp tool. And boy did it do the job. With a reach of up to 15m this was our secret weapon. No ladder or lift required.

So there you have it. Five things we learned from one of the largest beacon deployments in Australia. If you have any questions feel free to get in touch with us at team@lighthouse.io

And as always, good luck on your beacon powered journey!

James Irving

I'm a member of the context driven school of testing, in particular rapid software testing, session based test management and BBST. I currently lead testing at Inlight, a digital agency in Melbourne.

Melbourne, Australia https://au.linkedin.com/in/jamesir