Why Reporting On Keyword Rankings Is A Problem (And How We Can Fix It)
The SEO industry has always been obsessed with keyword ranking. To be honest, we only have ourselves to blame. Some practitioners have long promised (and many still do) first place positions for the keywords of your choice, giving the impression that being anywhere outside of the top three is an indication of failure.
So who can blame businesses for showing an equally obsessive regard towards keyword rankings compared to their competitors? SEO is all about increasing visibility in search engines, right?
Not so long ago, some SEO agencies began to disregard keyword rankings, coming to the conclusion that they didn’t matter, or at least were too dangerous to report on as KPIs. This was further impacted by Google’s exclusion of keyword data in Google Analytics, meaning that all we had to go on was rankings, not traffic and conversions. Many agencies started marketing themselves as content marketers or inbound marketing specialists instead.
It’s funny how these things operate in cycles. The truth of the matter is this: keyword rankings matter. They mattered before, in the old days of SEO. They mattered after it became fashionable to not care about keywords anymore. And they matter now. But how much they matter and what we do with them is where there’s a need for further education.
I hope to address this and clarify what keyword rankings should mean to you.
What’s the problem with keyword rankings?
The best place to start is trying to understand why keyword rankings can be misleading.
They’re not the bottom line
Some people felt that the nail in the coffin for keyword reporting was 100% (not provided) in Google Analytics. To clarify, a couple of years ago Google decided that it no longer wanted to tell us what keywords people used to reach our websites in Google Analytics. They did this in the name of privacy. What this means to us is that when we get traffic coming from Google (or any search engine, for that matter), we don’t know what the searcher typed in to reach us.
When you’re reporting on individual keywords this becomes a problem. So we’re number 1 for “random keyword” – great! But are we actually receiving traffic from it? Are we getting any sales?
Rankings are not a bottom line KPI. Rankings alone do not mean anything. Not even traffic, alone, means anything. What matters is leads, sales, conversions, revenue, etc. The stuff that makes you money.
Keywords contribute towards this, but they’re not the bottom line.
Broken promises
It’s a conversation every SEO has had. The client is in and we’re talking results. “So… we’ll get a top-three ranking for our main keyword, yes?” I wish it were as simple as that.
In Google’s world, we cannot guarantee anything. We can make predictions, predictions based on the competition already ranking for the keyword in question. We can talk about time periods or estimated budgets required, but we can’t guarantee rankings for a keyword. Instead we should extend our vocabulary. We should be talking in terms of online visibility (more on a concept called keyword indexes later), and other KPIs that we can meaningfully measure progress against.
Language is rich
Occasionally there are single terms or small sets of terms that are used significantly more than any other. In reality, the words that searchers type into search engines can vary enormously, even when looking for the same information. You might be keeping watch on three keywords for a particular product but what about the other hundred variations you’re not tracking? Not only is there a limit on how many keywords you can track, you won’t even be able to predict certain variations until they’ve already been searched for, and who’s to say when it’ll be searched for again?
The best results are those gained from tracking a large set of words and monitoring the results in aggregate. It comes back to visibility and the need for a more sophisticated way of keyword tracking.
Personalisation
What you see in search results is nearly always tailored for you; Google, as it tracks more and gathers more information about us and our habits, increasingly offers personalisation in the hopes you’ll be more satisfied with the results you see. This personalisation can be influenced by a number of things:
Location – searches with local intent will always consider the location inferred by your IP address or GPS in the case of mobile search
Device – when searchers use their mobile device, search engines prefer to show websites which are mobile friendly and penalising those that are not
Search habits and interests – search engines learn what sites you frequent and what your interests are, serving up results that are more suited to you
There are certainly other ways we’re not aware of that Google is personalising what we see. When you track keywords, you’re monitoring their position in one largely unpersonalised situation; we can track keywords by pretending we are in a specific geographic location or are on a mobile device versus desktop, but we cannot mimic the other infinite possibilities when it comes to personalisation.
A better way to report: keyword indexes
Despite all of the things that make keyword rankings unreliable, we should still be tracking them. They give us a good measure, outside of organic search traffic, of how much progress we’re making in key areas. No, we cannot see what keywords are bringing in search traffic in analytics software anymore, but we can see in a general, unpersonalised way, whereabouts a keyword, or set of keywords, features in the rankings.
Rank indexes
A better way to report on keyword rankings is to use the idea of rank indexes. The concept is fairly simple. Instead of reporting on the rank of individual keywords, we take a set of keywords that all represent a type of query class and measure the aggregate rank of the group. The query class will just be a set of keywords with something in common, perhaps related to a single product, service, or idea. If you have an ecommerce store, for example, you’ll probably have a query class for every individual product type.
By doing this, we’re reporting only on a single number – the average rank of a bucket of similar keywords. How this single number varies tells us if we’re moving in the right direction or not. This is a reporting method we are beginning to introduce at Zest, one which will make our reports more focused, meaningful, and valuable to our clients.
To give you an anonymised example, you might see a graph like the following (the lower, the better, by the way):
keyword rankings graph
Rankings are not the be-and-end-all measures of SEO progress. They are indicators and are of more use to us, the people doing the SEO, than they are to the client, because rankings must be considered within the context of an often complex whole. This doesn’t mean that we shouldn’t be reporting on keywords at all, however. There’s just a better way.
Rank indexes are easier to read, make more sense to the client, and go a long way towards alleviating some of the problems inherent to keyword tracking described in this post.
How are you measuring your keywords?
