Google’s Ben Gomes has posted about the future of search. Nothing in it surprises me. Quick summary:

• From answers to journeys: search to be more personalized and contextual, helping you “resume tasks where you left off”
• Queryless information: surfacing information “relevant to your interests” without you asking.
• More visual results. Because everyone likes a picture.

Personally I would prefer search to be improved in different ways. I would like:

– clearer separation of ads from search results. It is to my mind wrong that brands have to advertise based on their own brand name, just to ensure that users searching for their brand find the official site, and not a competitor or intermediary

– Better results. As a techie I am often looking for answers to technical queries. Search is very useful, but in general, I find too many results with the same question but no answers, too many old results that are no longer relevant, and not enough focus on community forums (where the answers often exist).

– Better authority. As a journalist, authority really matters; and I do not mean “reported by a well known news source”. Authority means first-party information, the announcement from the actual people or companies involved, the information on the first-party sites or from actual employees. Finding this is quite a lot of work, and the algorithms could be much better.

What I do not want includes:

– over-personalized results. There are two reasons. First, I am wary about giving away all the personal data which Google wants to use to personalize results. Second, factors like objectivity, balance, and accuracy matter much more to me. I do not want my own version of fake news, results designed to please me rather than to inform me. Nor do I want this for others, who may end up with a distorted view of the world.

Of course it depends what sort of search you are making. If you search for “best restaurant in Oxford”, what do you want? The most highly-rated restaurant (by some standard) among places where you typically choose to dine? Or the best according to the general population? Or the best according to top restaurant critics? It is not clear; and a journalist (say) might want a different answer to someone looking for a place to eat tonight.

All of this touches on a key point, which is search results versus marketing. Is search a way of researching information on the internet, or a marketing tool? I want the former; but unfortunately it will always be, at least in part, the latter. Particularly as we are unwilling to pay for it.

– too few results. Ten blue links was a luxury: 10 answers to the same question, hopefully from different sources, so we can see any diversity and make a selection. The search, um, experience now more often gives us just one result, or at least, one prominent result and more available if you work at it. This is especially true of voice assistants as I’ve noted elsewhere. There are obvious risks in the trend towards one-result searches, including dominance of a few sources (and the squeezing out of the rest).

– opaque results. Wouldn’t it be great if you could find out why, exactly, Google has chosen to give you the results it has yielded. Puzzling this out is of course the realm of countless SEO experts, and there is always the argument that if too much is known about the algorithms, they are easier to game.

The downside though is that we have to trust Google (as the dominant provider) to do the right thing in many different ways. It will not always do the right thing. If its vision of the next 20 years of search is accurate, we are being asked to become increasingly trusting, even as we are also discovering, through devastating political outcomes, that you cannot trust big algorithm-based, commercial internet providers to look after our best interests.