Improving our sites when keyword data is hidden

[smallcompany]

With Google's search market dominance combined with people that are logged into Google account, more and more keywords are being hidden to our tracking/analytic software.

Turning to improving our sites without specific SEO (keyword) focus is fine with me, but not knowing what people searched for and which keywords converted and which were a miss is at least very strange.

For example, my keyword No.1 for one of my sites produced 1536 visits for the last 30 days, and the second one called "keyword not defined" caused 1470 visits. Now, I know that in some cases keyword is not available due to other causes, but the majority now is that Google privacy thing. And it's growing...

I wonder if anyone already stopped looking into keywords and keyword oriented optimization?

[lucy24]

There's a big unanswered question. Is it random?

If any given search term is equally likely to be hidden, then you've still got the same information. Just less of it.

But if some search terms-- among searches leading to the same site-- are more likely than others to be hidden...

I think most users don't know that logging in hides their searches. So you're not missing the users who are most interested in concealment. You may be missing the ones least interested in concealment. The ones who sign up for everything and log in everywhere. Are users with concealed search terms more likely to be on Facebook? More likely to post to blogs?

There's a ###load of information going missing. Way beyond "What did they search for?"

[Robert Charlton]

There's a big unanswered question. Is it random?

If any given search term is equally likely to be hidden, then you've still got the same information. Just less of it.

I think the information is likely to be the same until something that we're not yet seeing changes the nature of the hidden searches... ie, how something in the searching situation changes how searchers behave.

The largest such changes, I'm thinking, are likely to be in mobile, which is the area of largest anticipated growth. Mobile searches are secure and won't return referrers. Mobile interfaces are rapidly changing, and this might also affect the nature of searches.

Personalization conceivably will also change searching patterns by changing the results that searchers have seen before new searches.

The largest personalization factor I've thus far seen, btw, even on desktop, is location. More often than not, location is not entered into queries when localized results are desired.

On mobile, it's likely that results, which are in many cases already hyper-localized, are likely to become ever more narrowed down by place, and conceivably by time... and by intention as perceived by time and place. Eg, a mobile search from a downtown area at dinner time for a restaurant is likely to narrow down the geo-area quite a bit, with no location information entered in manually.

Conceivably, Google... or sites returned by Google... might even go into a restaurant app-like mod and offer a menu-driven dialogue to rapidly narrow down choices. Assuming that what we call "websites" receive any of that traffic, it's hard to say what the stats might look like.

I can envision similar localization for some types of shopping.