I was poking around a bit today with identifying mobile search traffic. I dumped a few million records out of the Mowser traffic DB and parsed through the referer URLs to pull out keyword info from searches that land on Mowser pages. First things first, I wanted to share. So here’s my current version of a mobile search term referer parser. I just wanted to get that out of the way cause I’m starting to get tired, by tomorrow I might forget about what I did if I don’t post it.

There’s a lot that’s different between the mobile world and the web world currently (and it’s all headed for what’s sure to be a noisy, violent, and I’m betting pretty damn interesting collision). In the online world Google owns most of the search market, for better or for worse. However on the mobile side many service providers are building out their own search index, usually licensing some technology from someone to run their own site or white labeled version. Say what you want about Google, but at least as a property owner I know how to get search traffic to my site: get ranked on Google. What does the world look like off-deck if every operator keeps building out their own search index? How does it affect end user perception of search as a service as well if every service they try out works different?

Also, how as a site owner do you attempt to tune your site for search traffic when many of the search services that are driving users to your site aren’t available unless you’re a customer of the same carrier? Check out these referers I have from Mowser:

If I can’t even see the pages it really makes it difficult for me to understand traffic flow to my site and attempt to tailor my content appropriately. Carrier searches are actually driving a pretty decent amount of traffic still, but already they’re lagging behind Google in terms of driving numbers. Carrier search is only going to get worse by comparison, especially as higher capability devices work their way out into the market and folks begin expecting their mobile search to work like their desktop search (because that’s the way the rest of the web does).

This is just another one of those issues where we’re probably going to have to suffer through long and painful process of carriers finally coming to understand search after multiple failed and user-frustrating attempts. I’m not saying that mobile search isn’t different than web search. But if you’re going to run a web service run it out on the web. Pay attention to all the people who are affected (and who can help you improve your service). Operators: going off and running web search with everyone having their own instance off in their own little tidepool is stupid. Please stop doing it.