The first response that John gave was:
I found this conversation particularly interesting because, while John Mueller knows a lot about SEO (he’s an authority in the space from Google) he advised someone to “test it with a small part of your site & see how it works out with your site & your users”.
Lesson: Test, Try, Experiment
John Mueller is an expert in how Google works and his advice to someone asking questions is “test it… see how it works” — I wouldn’t take that to mean that John doesn’t have any insight on what the possible results may be. I’m sure as someone who understands the algorithm quite well he could make educated predictions about what the results would be. However his response was very appropriate for a few reasons — there are several major ways we could look at this type of response (which is typical) from someone who can speak “more formally” for Google. They speak in a qualified manner for a few reasons:
- They don’t want to give too much “secret sauce” away. (My personal conviction is that they are generally nice guys who tell us all they can reasonably tell us — I’m not cynical about them hiding stuff)
- They don’t say “this is definitely going to work” for legal reasons. (Definitely that’s a part of the picture)
But my favourite and IMO the key takeaway and most accurate reason is this:
- Because I come from a technical background I know that, when it comes to technical stuff, especially highly complex things like the Google algorithm — highly deterministic predictions are simply not possible.
In other words, and this is absolutely not an insult to John, it’s likely that even he doesn’t know exactly what would happen if a site were to write unique product descriptions.
The reason is that Search Engines are so complex and rely on such a large variety of signals that saying something like “yes, writing unique product descriptions will make you rank better” would just be silly. There are so many other factors. Your competitors might have a stronger backlink profile. Your site might have crawlability issues. Your manufacturer(s) may not have published the original product descriptions online, etc.
There are so many things that could influence the outcome of such an exercise that it’d be unqualified to just say “Hey, do X and Y will happen” — hence the need for empirical evidence and experimentation.
Lean Startup Thinking
It is this highly dynamic environment — with many factors influencing the final result — that makes SEO lend itself so well to “Lean Startup Thinking”
The Lean Startup is a book by Eric Ries for entrepreneurs and startups. One of the fundamental takeaways of the book, in so many words, is this:
- We tend to make leap-of-faith assumptions in life and business (addition: and SEO)
- We think that X will lead to Y, and thus sell ourselves an oversimplified model of reality (which impacts how we do business (and addition:SEO))
- We should question those assumptions before sinking large amounts of money into something that might potentially not work.
- We should make learning occur as fast as possible; fail fast, fail often.
- Small, incremental failures / successes teach us lessons and present us with empirical evidence that a certain course of action does / does not work, and help us to figure out why.
All of this applies to SEO. As SEO people we know a lot about how Google works, and what has worked in the past. But Google is constantly changing and no two campaigns are exactly the same — every client or website that we look at is unique, they have unique competitors in a different niche from the previous ones we’ve done before. So while many ideas carry across various niches, not all do.
And that’s why we need to:
“test it… [and] see how it works”