Trust But Verify

I read a piece in Business Wire with great interest on the entry of the Broadband Information Services Consortium (BISC) into the broadband mapping fray. After all the heat that Connected Nation has gotten in recent months for being the paid handmaiden of industry, I'm curious to see what a large-scale alternative might look like and how to build a better mousetrap.

I hasten to add that mapping is crucial. But as I've written it is only the first (or perhaps second) step in improving broadband access. While several states have contracted with Connected Nation for mapping services alone, I'd wager the greatest results in broadband deployment are those that take advantage of the broader community mobilization such as that which has occurred in Kentucky and several other states. That is because a map is necessary to point out gaps in service. But those gaps exist because large provider don't see near-term opportunities to recoup their investment. So mobilizing local knowledge and forging local coalitions is essential if anything is going to change in those areas.

And that's where the real power of Connected Nation's approach lies.

So what does BISC offer? It's really hard to tell. Here are a few clues from the Biz Wire piece:

"...provides states with customized solutions to broadband mapping to address the full supply-and-demand broadband continuum..."

"... ensures the most accurate, fully verified and up-to-date information available for broadband mapping..."

"...“Our collective experience and platform enable us to compile the multiple layers of real-time data of location and serviceability, either as a full-service approach or as a complement to state efforts....”

"Broadband maps created with geographic information system (GIS) technology provide an advantage to states..."

These notions all sound great. But one question remains: where do the data that form the base map actually come from? The Biz Wire piece doesn't say.

Connected Nation has been roundly criticized for accepting as gospel the information that they receive from broadband providers. That's a fair line of attack. Why should we trust providers, particularly when the data they release comes with strings attached?

As I've written previously, while they've not called it "crowdsourcing" in the past (what not hip enough, dudes?), Connected Nation has encouraged public validation of its base map. The information that providers contribute is not cast in stone, but is viewed as a starting point. Connected Nation has provided a mechanism for consumers to question the accuracy of the map, soliciting inputs from the public as means of improving on what providers are willing to share. For example, in Kentucky alone, over 4000 inquiries have been made based on the data providers have made, improving the quality and accuracy of the map as a whole.

The BISC plan seems to make heavy use of crowdsourcing (which amounts to polling or surveying the public) and extrapolating from the poll to estimate conditions over all. This is apparently the sum total of the approach taken by Broadband Census, which begins with a blank slate and maps on the basis of voluntary polling (not sure how many data points the Census has for the whole nation, but I've been told it's fewer than the 4000 Connected Nation has for Kentucky alone (as a corrective, I should, add, to a base map drawn from provider data)). As any statistician will explain, the fewer data points you have, the larger the error terms (i.e., the more the inaccuracy). So if you start from a blank slate it takes a lot of data points to generate anything meaningful.

Connected Nation's maps are not perfect. Nor are they intended to be. Nor could they be. Broadband deployment changes rapidly. And providers (for all sorts of reasons, and not just the big ones) don't provide entirely accurate spatial depictions of their deployments.

The big question is: Where does the map start? From a blank slate (in which case crowdsourcing is likely to contain lots of inaccuracy)? Or from a flawed, incomplete, temporally-bounded map of provider data (in which case crowdsourcing can and, in Connected Nation's case, already does make a big difference)?

Although they claim to be operating transparently, we don't yet know where BISC's base map comes from. I'll withhold judgment until I do.