Here's What Hourly Google Search Result Tracking Looks Like

24 hours in the life of a search result

As an ancient Greek man once said, "the only thing that is constant is change." I found out just how true that is of Google's search results by tracking the position of the top 100 pages for six different searches every hour over a 24 hour period.

The results are beautiful.

The Competitive SERP: [car insurance]

At least in the UK, [car insurance] is a competitive, high-value search result with around 500,000 searches every month. Here, the top 40 desktop results remained pretty static throughout the day, and those pages finish the day in pretty much the same positions aside from some minor fluctuations, something impossible to spot with daily rank tracking.

After position 40 things start to get a bit hectic, sometimes with wild swings of 50 positions over an hour. Throughout this post, the more white background you see in the charts the more that URLs we're tracking have disappeared from the top 100, leaving a gap. That's a strong indicator of a turbulent SERP.

The News SERP: [donald trump]

This data happened to be recorded on the day Donald Trump visited India. It's chaos, but probably what you'd expect from an changing search like [donald trump], where every hour new articles are published, updated or deleted. It's important to know that the data in these charts exclude search result 'features' like Twitter pull-outs and Top Stories news boxes.

(Weirdly, on desktop Donald Trump's Twitter account remained rock solid at #97 throughout the 24 hours.)

The Local SERP: [plumbers near me]

How many new plumber pages can there be in any one day? I thought this would be a reasonably calm search result with some flux here and there, but it’s much choppier than that. As explained above, these results exclude search result ‘features’ like Twitter pull-outs and – importantly here – the Local Map Pack.

But still, in any local search result, the ‘normal’ results are still heavily localised.

The Spammy SERP: [payday loans]

Of all the tracked keywords I was convinced that this would be the most turbulent SERP of the study – but that wasn’t the case at all. The keyword [payday loans] is a profitable one, and has a long history of being one of the most spammy niches around. Google even introduced a specific payday loans algorithm back in 2013 to combat the problem.

Back then, fly-by-night sites would pop up on the first page for a few days then disappear without a trace when Google eventually caught up. The SERP today looks a lot more stable however, which might suggest they’re now on top of the problem. Google is likely holding these results to a higher standard than most, which makes sense considering how they think about Your Money or Your Life (YMYL) queries.

The (other) Spammy SERP: [voucher codes]

There’s also a lot of money to be made from voucher code searches. As you’ll probably know if you’ve ever looked for discount codes during checkout, the results can be very hit or miss. Plenty of sites will reel you in with outdated codes (or no codes at all) so that they can cookie you and score some commission, so that might shed some light on the type of players in this SERP.

The site: SERP: [site:bbc.co.uk]

Out of curiosity, I wondered if a [site:bbc.co.uk] search would change hour-to-hour. In Google, a [site:domain] command shows results from a single domain – and the general SEO wisdom is that results are broadly ordered by link authority. That is, the most internally and externally linked pages on the site will show up first since they’re probably the most valuable and important on the site.

That doesn’t look to be the case here, but in hindsight, I probably shouldn’t have chosen a site that publishes hundreds of new pages a day which is probably throwing these results off quite a bit… ¯\_(ツ)_/¯

Wrap up

There’s not a ton of actionable insight to be had here. If you’re an SEO, you probably can’t switch your rank tracking to hourly, and even if you could what would you do with all that data? What it does do, is make clear that if you’re trying to rank for a keyword, then you can’t easily judge your progress on a linear scale.

Position 50 this morning may well be position 20 before you go to bed, but once you’re in the top 20 you can probably expect much slower movement to the top. The reverse is also true of course – the decline in ranking we saw was a page that dropped from position 1 to 100 over the course of an hour (voucher codes, desktop).

And credit goes to Simon and Chris from Strategiq, who were first to do this a couple of years ago. You should check out their write-up, too.

About the data

If you’re curious about the details of how we did this, here you go:

• Results were pulled on Monday 24th February 2020, (which happens to be the 9th anniversary of the original Panda roll-out).
• As far as I know, hourly tracking isn’t available with any of the popular commercial tools so we had to roll our own.
• 100 results per keyword were pulled using the &num=100 Google search parameter, rather than 10 at a time.
• The search location for was set to the iProspect Manchester office address in Manchester, UK.
• We excluded everything that wasn’t a ‘normal’ result. That means we ignored Top Stories boxes, the Local Map Pack, Image and Videos results etc. That’s less reflective of real life but helps make the comparison between search results fair.