SERP Volatility and Freshness Signals

Full transcript

989 words · script draft, under review.

Intro

You analysed a results page a month ago. You scored it, you found a weak spot, you planned a page. Today you look again, and half of the top ten is different.

Was your analysis wrong? No. The results page moved. Some searches are rock solid for years. Others reshuffle every week. And some only reward pages that were published very recently.

If you don't know which kind you're dealing with, every SERP analysis is just a single photo of a moving target.

In this episode, I'll show you how to measure that movement: how to take snapshots, a simple rank churn metric, and the signals that tell you a search wants fresh content.

Two different questions

Let's separate two ideas that often get mixed up.

Volatility is how much the results page changes over time. Which pages are in the top ten, and in what order.

Freshness is whether Google prefers recently published or recently updated content for this search. A search can be volatile without being about freshness, for example when several similar pages keep swapping places. And a search can be about freshness with fairly stable domains, where the same news sites always rank, but with new articles every day.

Both matter for planning. A volatile results page means your position will move too, so you need more measurements before you judge it. A freshness-driven page means that publishing once isn't enough. You're signing up to keep updating.

Taking snapshots

You can't measure change from one look. You need snapshots.

A snapshot is simply the top results for a keyword on a given date, stored as rows: keyword, date, position, URL, domain, and result type. Same location, same language, same device every time. If you change the settings between snapshots, you're measuring your settings, not Google.

How often? For most keyword sets, once a week is enough. Daily snapshots mostly add noise and cost. For a hundred keywords at depth ten, a weekly live SERP pull is roughly thirty cents a week with DataForSEO. That's about a dollar and a quarter a month.

Keep the raw response too, not just your parsed rows. Then you can go back and extract something you didn't think of the first time.

And add a small control set: a handful of keywords outside your niche that you track the same way. If everything moves at once, including the control set, it's probably a broad Google update, not something specific to your topic.

Measuring rank churn

Now the metric. I use two simple numbers per keyword, between two snapshots.

The first is turnover. How many of the top ten URLs are new compared to the last snapshot? Three new URLs means a turnover of thirty percent. It's easy to explain, and it catches the big movements.

The second is average movement. For URLs that are in both snapshots, how many positions did they move on average, up or down? A URL that dropped out of the top ten counts as having moved to position eleven. This catches reshuffling, where the same pages just swap places.

Then you look at it over several weeks. One turbulent week is an event. Four turbulent weeks in a row is the nature of that search.

Here's what that looks like with made-up numbers. Keyword one, "hand coffee grinder", has almost no turnover for eight weeks. Keyword two has one big spike, then calms down. That spike lines up with the control set, so it was a broad update. Keyword three, about new espresso machine releases, has high turnover nearly every week. That's the nature of that search.

Freshness signals

Freshness has its own signals, and you can see most of them in a single snapshot.

Dates in snippets. If most of the top results show a recent date, Google is surfacing recency.

The current year in titles. When many titles say "this year's best", or the actual year, searchers expect up-to-date content.

News-style features. A top stories box or a news block means the search is partly treated as news.

And the age of the ranking pages. Over a few snapshots, you can see whether new URLs keep entering at the top, or whether old ones hold.

When three or four of these line up, I label the keyword as freshness-driven. That label goes straight into planning. A freshness-driven keyword needs an update routine, not just a launch.

Using the labels

So what do you do with these labels? Each one changes the plan in a different way.

Stable keywords are the easiest to plan for. Build once, check the results page every few months, and judge your position on a normal timeline.

Volatile keywords need patience in measurement. When you publish a page for one of them, your own position will jump around too. Don't judge it on one good week or one bad week. That matters a lot when we get to re-measuring in episode forty-six.

Freshness-driven keywords carry an ongoing cost. Someone has to update the page, every few months or every few weeks. That cost belongs in your prioritization, next to the expected traffic, and we'll use it in episode forty-four.

Common pitfalls

Three pitfalls.

One: comparing snapshots taken with different settings. Different location, device, or depth gives different results. Lock your parameters and store them with every snapshot.

Two: depth too shallow. If you only pull the top ten, you can't see a page climbing from position fifteen. For keywords you really care about, depth thirty costs about twice as much and shows you who is coming.

Three: reading meaning into one week. Results pages wobble. Judge on several weeks, and always next to your control set.

Wrap-up

So: weekly snapshots with fixed settings, turnover and average movement as your churn metrics, a control set, and freshness signals as a label. The snapshot table schema and churn script are linked in the description. Next episode: the same search, but on mobile, on desktop, and in different locations, and when those differences actually matter. See you there.