The Future of Search: A One-Act Play in Three Acts (Act Two)

SEW presents the weekend column. We have reserved Saturdays to invite people outside of ‘the industry’ to share their thoughts on our industry. For three weeks Sci Fi writer, Laszlo Xalieri (nom de plume), is sharing his opinion on the future of search engines. Enjoy Part 2. (read Part 1 here)

Targeted marketing is awesome. Keep in mind that I say this as someone who is being marketed to as opposed to a marketer. I know that by handing over a smidge or two of information about myself, I participate in a process that cuts down on waste.

Some people in the market are offended by this. I don’t get it. I’m doing marketers a big favor by allowing them to reduce costs across the board — less paper and ink and postage in traditional junk mail (still a big business in some sectors), and less bandwidth and server cycles wasted showing me opportunities that aren’t relevant to my tastes/gender/species/buying habits.

By giving up a little information about myself I help marketers reduce costs tremendously. In exchange, however — and this is especially important online — they don’t waste my time and peripheral attention by showing me stuff that isn’t entirely for me. (This applies to SEO-tuned search results as well as to the more in your face PPC ads.)

I’m still waiting for this to happen for cable television. I’m nearly offended that my integrated cable and Internet provider isn’t sniffing my household Internet traffic to show me more relevant :30 spots in the commercial breaks.

Surely Comcast, recently determined to be the Worst Company in America 2010 by The Consumerist, could be evil enough to do me the favor of looking at keywords in pages and content/context-aware ads I’m already being served (so they don’t even have to do the work themselves!) and swap out those spots for feminine hygiene products and cars I can’t afford for those private health insurance and credit counseling agency ads that are much more relevant to my life. It truly amazes me that they don’t do this. They still pay Nielsen for broad demographic data instead of, you know, selling it to them.

Image from Consumerist

Because the more you know about who you’re marketing to, the more likely advertisers will get a sale and the more valuable your service is. You can charge your advertisers more. And they’ll happily pay the extra because they’ll be making more money. It’s easy math for everyone: the advertisers, the media channel, and the consumer.

Well, at least until the marketing data gets stolen by identity thieves, purchased by unprincipled third-party marketing fiends, or subpoenaed by Homeland Security so they can correlate potential criminal and/or terrorist activity to little details like who views pages that feature bulk fertilizer ads.

Don’t act like you don’t know that certain people and organizations aren’t already hell bent on abusing this data if they can get their hands on it.

There are a zillion people out there who still haven’t caught onto the difference between correlation and cause — marketers included — and are willing to waste a lot of resources and consumer (and/or potential perpetrator) goodwill by letting huge amounts of data fall into the wrong hands, sometimes even in exchange for a stack of cash. Or just as a favor.

AT&T turned over call history data to federal terrorist hunters without a warrant and was, in my book, insufficiently raked over the coals for it. The Bush II regime was happy to sign a law requiring libraries and bookstores to submit lists of names of people who bought or checked out books on particular topics, complete with an automatic gag order that prevented anyone discussing the order to hand over the data. I shut down a novel I was writing because it was suddenly an inconvenient time to research historical Islam for background material, especially when it was no secret I had many friends overseas.

The proper term is “chilling effect.” You trust the people who have your data to not sell you down the river.

Consumer/user privacy concerns are no joke, as Google and Facebook well know. Your “Don’t Be Evil” policies are only as good as your data security, your ability to extrapolate the consequences of your actions, and your immunity to the temptation of a tasty stack of cash. Oh, and your interpretation of evil.

Sniffing traffic of unsecured networks as you drive past, Google Street View? Why not peek in people’s windows and flip through the envelopes in their mailboxes too? Why not check to see if people left their front doors unlocked and do a little filming from the foyer?

Just because you can doesn’t mean it isn’t evil. It’s all a matter of trust, and not betraying it. Betraying trust is evil — and a poor long-term business strategy.

The only solution is to give every user on the Internet one or more anonymous tokens that gets tracked wherever they go, some pretty ironclad opt-in policies for how you’d like your history to be used, and a system that only allows a token to be connected to a physical identity by a warrant signed by a judge. Right now we’re faking all of that with (overly identifying) browser strings, (never-expiring) cookies, and a hodge-podge of (overlapping, contradictory) laws squatted out by legislators who grew up when airplanes were made of wood and entire neighborhoods shared a single “party-line” four-digit telephone number.

So that’s the present: riding that line between downright aggressive invasion of privacy and maintaining security of what you’ve been given trusted access to. What’s the future?

Well, Google is certainly interested in the future. They’re trying to buy a peek. I’m fascinated by the concept of topic momentum, where events in the news and/or popular consensus reality are treated like a note in a symphony, charted more or less in terms of attack, sustain, decay, and release.

That really is only half the threads in the fabric, however. You still have the completely unpredictable elements of actual events to deal with. In any case, though, this should allow them to speed up access to predicted popular topics by prioritizing the caching of related material. And then they can shuffle the priorities when they see a spike that signals the “attack” phase of a new note.

Perhaps they can combine this attempt to grasp the future, their knowledge of who you are, and IBM’s predictive analytics (for predicting juvenile crimes) to offer, instead of their current autocomplete bizarreness and “popular reading” suggestions, a tailored suggestion box based on knowledge of the sorts of topics they’ve known you to search for in the past?

(I should point out that a popular criterion for sentience in an organism, among cognitive scientists, is the ability to anticipate. But I digress.)

Nearly coincidentally, Twitter has built this predictive engine by sheer accident. More on that in the final act. Stay tuned!

Related reading

Gillette video search trends
serps of tomorrow
2018 SERP changes impact SEO