Return to Articles 6 mins read

SEO is not zero sum - SearchPilot

Posted November 25, 2021 by Will Critchlow
Last updated August 7, 2023

I have regularly heard people say that SEO is zero sum and it seems that almost half the industry may think that’s true:

“Zero sum” is a concept from game theory meaning “a situation in which one person or group can win something only by causing another person or group to lose it”. What I believe people mean by it in an SEO context is that every time you move up in rankings for a particular search phrase, other websites necessarily must move down to make space for you.

This is intuitive, but I believe it’s wrong.

It’s wrong in at least two ways:

1. SEO isn’t zero sum in rankings

There are searches that return very few results (anyone remember googlewhacking?), and so it is possible to do work that creates a new result on a search results page without displacing anyone else by adding to the end of the list of results. Technically this is possible for any search result since no searches have infinite results.

This is true, but it’s not very interesting, and it’s kind of an edge case.

The slightly more interesting version of this is deliberately creating content about new things: every day there are combinations of words that have never been interesting before that become new legitimate search terms. Someone has to be first to write about them. Maybe it’s you?

(The broader case of creating the conditions for new content to be created at scale is covered below).

2. Rankings aren’t the right way to think about the game

The more interesting argument is that rankings aren’t the right way to think about the benefits of SEO (for very similar reasons to the reason that we at SearchPilot don’t measure the results of our tests with rankings). Rankings don’t pay the bills; what matters is the qualified visitors that you drive to your site, and when we are talking about traffic, SEO is definitely not zero sum.

If we’ve learned one thing from our SEO test case studies, it’s the power of clickthrough rate. Many of our biggest uplifts (and, ahem, our biggest losses) have come from changes that have the potential to affect clickthrough rate - whether in the title, meta description, or structured data / rich snippets. When you take into account the fact that different pages ranking in the same position for the same query can have radically different clickthrough rates, you might start to realise where I’m going with this line of argument.

The short version is that, to the extent that you either promote pages that are better, make pages that are better, or make pages better, your work can improve the whole search results in such a way that the total number of satisfied users goes up.

Said another way: there is no particular reason to think that doing something like making your title more compelling only extracts clicks from other results - I would in general expect it to extract some of those and encourage some previous “no click” searchers to click on your result.

In the extreme case, where nothing good was ranking previously, and most users were giving up without clicking, creating a good page that ranks well could capture a large number of new clicks. For the purposes of this analysis, I’m not splitting out clicks to Google-owned properties, and clicks on adverts. There are various ways you could treat them (counting those as “no click” on organic results, or considering all clicks to any link on the page), but none of these change the conclusions.

As a simplified worked example, consider your page (A) and your competitor’s page (B) on a single search result where each user only clicks on one result:

  • Imagine before your changes that you (A) rank in position 3 with a snippet that is so compelling that for any given position, it generates 1.5x the clicks of the page ranking in position 2 (B
  • Suppose position 2 generates 2x the clicks of position 3 (a reasonable assumption)

To make it concrete, that might look like:

  • Page A
    • When in position 2: 1,476 clicks / month
    • When in position 3: 738 clicks / month
  • Page B
    • When in position 2: 984 clicks / month
    • When in position 3: 492 clicks / month

You do some work, and A moves up to position 2, while B falls to position 3:

  • Before, there were a total of 1,722 clicks / month to A and B combined
  • After, there are 1,968 clicks / month to A and B combined - an increase of 14% just to those two pages

These effects are magnified when we create great new content. It can have all of the numerical effects explained above and if we create genuine value with content that is better than anything that previously existed, it’s fairly easy to see that user satisfaction goes up because not only are more people clicking through, but users are more satisfied on average. When you consider that a lot of enterprise-scale SEO is about systems and processes, and you realise that some of the biggest wins have been huge wins for the web as a whole, you can see how this scales up. [This whole post was at least partly inspired by that comment on Hacker News by @Patio11].

This focus on user satisfaction as a metric leads into my final point.

Going all the way down the rabbit hole

If we believe that, over the long run, Google iterates and converges on satisfying user intent, the effects outlined above become even stronger, because not only are pages likely to get more clicks per position as they are improved and move up the rankings, but each click is also more likely to result in a satisfied searcher.

[This concept is somewhat similar to the way consumer surplus can increase through trade in economics.]

The exercise of showing that the economic advantage resulting from being the truly most satisfying result should result in the company delivering the greatest consumer surplus being the most likely to improve their own rankings is left as an exercise for the interested reader.

Reader comments

Manley, CEO at Corigan kindly volunteered to read an early draft of this post and contributed an additional point that I hadn’t thought of, which I have paraphrased here:

SEO makes people search more

You can argue that in aggregate, the process of SEO when done across the web has additional positive externalities. Making good content and making sure that Google can understand what search queries your content relates to improves the experience for all searchers. The ultimate result of making search engines work better, and improving the experience of searchers is searchers searching more. This means there will be more clicks to go around.

Summing up

That was a great deal of talk of clickthrough rates and economics, but I think the general principle can be summed up as follows:

Either the website / web page is the experience, and you are making the actual experience better, or there is some underlying good thing that you are making more discoverable (whether it’s a product for sale or an experience to enjoy). Making better experiences more discoverable at the cost of less discoverability of worse experiences is a net win, so all we need to believe to be convinced that SEO is positive sum is that Google is good at their job of ranking the better experience more highly.

Image credit: Piret Ilver.