Can they take a great selfie at the park? Yes. Capture a stunning 360-degree photo over the city? Yes. Follow your epic dirt bike trip? Yes. Search for a lost hiker? Scan your construction site? Survey a bridge? Digitize your cell tower? Yes, yes, yes – to everything yes. Really? Is there anything an off-the-shelf drone can’t do?

This is the current state of the drone industry. A million affordable, capable aircrafts sitting on shelfs with hundreds of companies looking for problems for them to solve. The potential applications of drones appear limitless, but for all the hype and investment, drones have failed to live up to forecasts. Why?

There is one artificial constraint that all off-the-shelf drones share – low fidelity sensors. The drone industry has made steady progress over the years, and aircraft hardware, autonomous flight software, and data processing and analytics engines have all improved. But evolution has been bottlenecked by the capabilities of the sensor and the data they can collect.

Off-the-shelf drones weren’t built to capture commercial data for enterprises, and in no case is this clearer than in telecom. They spun out of hobbyists who wanted a better remote-controlled flying gadget that could take pictures and stream video. Then they were adapted by startups to fit the commercial use case. Companies exploded on the scene with new and novel ways industries could use these off-the-shelf drones. But that was the problem. We tried to take a hobbyist toy and convert it into an enterprise tool. We were given a hammer and told to go find nails. And today, while the industry has made advances in autonomous flight, 3d data processing and AI-based analytics – very little has evolved in the data quality itself.

Great product categories start with a need first. They solve for a “missing” in a given market or industry. So, what’s the missing in tower industry?

Yes, tower inspections are manual, dangerous, costly… blah blah bah – you’ve heard this pitch a thousand times. I’m guilty of evangelizing this message myself. And while interested, towercos aren’t seeing the kind of ROI here that would justify implementing a program at scale. Why not?

Tower inspections are not the problem, they are simply a symptom of a bigger problem. Legacy data is the underlying issue. Towercos have inadequate tower data – meaning it is incomplete, inaccurate or missing altogether. This creates problems for Finance to accurately capture billable revenue, Sales to effectively communicate available inventory, Engineering to calculate structural capacity, M&A to validate and onboard new sites, and yes – Field Operations who must make redundant site visits and conduct manual inspections.

Comparing the Visual Intelligence (MACS-3D) sensor with a DJI P4P

Video 1: A side-by-side comparison of 3d models captured on the same tower.

If you already have high quality legacy data, then perhaps data from an off-the-shelf drone can be used to make some assumptions and educated guesses. But after decades of tower data changing hands between companies, service providers and personnel, I’m willing to bet there’s not a towerco in the world who is 100% confident in their legacy data. This is why we see so many truck rolls and a heavy reliance on third-party providers.

Drone data cannot solve for the legacy data problem until it can achieve millimeter accuracy and generate engineering-grade 3d models. Attributes like lattice structure width, pipe thickness, bolt sizes and the degree of twist and sway are all essential measurements that tower technicians must collect. To augment and replace measurements taken in the field, the 3d digital model generated from a drone must achieve no less than this level of detail.

Analytics engines must recognize the make and model of inventory, not just detect the shape of antennas, dishes and RRUs. They must produce engineer-stamped 2D and 3D CAD drawings, not just tell you the estimated heights, tilts and azimuths of equipment. They must automatically generate reports, and not require a tower climb to validate measurements.

When all you have is a hammer, everything starts to look like a nail. The drone industry has simplified the telecom problem to “inspections”, and formulated a solution around precisely the tool they had in hand – off-the-shelf drones (the hammer). However, to solve for legacy data, the telecom industry needs a different kind of tool. Instead of adapting problems to the tool, it’s time we adapt the tool to the problem. In this case, it’s time for a new kind of drone sensor.