Hourly Rank Tracking – One Year On
Search Engine Optimisation

Hourly Rank Tracking – One Year On

16th May 2018

A year ago I carelessly tweeted an image of what the top 100 search results looked like when tracked hourly. We thought it was ‘kinda cool’, and it turns out a few other people did too. I recently had the opportunity to talk about it again at SearchLDN, and for those who missed it here’s a write up (slides at end too).

Where it Started

A scene most of you might be familiar with; you all sit around a table with a client/stakeholder following a period of ranking fluctuations. That’s where this started, it was quite stressful, a money term had dropped 4/5 positions and it was hurting. During the meeting the positions were fluctuating; whereas the ranking reports were relatively steady – daily tracking wasn’t doing justice to how volatile things were.

This turned into a “what if” conversation with Simon: if we tracked rankings hourly, how different would it look? Luckily Simon had just built his own rank tracker (because, why not?) so we were in the really fortunate position to answer that question.

The first hourly rank tracking screenshot

If you want to read any more on how things started, check here, but the below image is a dramatisation of about a year’s work:

A dramatised history of the development of the hourly tool via WhatsApp

Whilst I run the risk of severely under-selling the amount of hours that we have poured into this, we now have an hourly tracking service, which can scale and has a UI.

Closed Testing

In creating something a little more robust and user friendly, we had a great opportunity – to invite others to join in. Over the last six months we have invited a very limited number of people who have said they were interested to test it and see what they could find/discover. Oh, and this is all for free. Why’s that?

Why we're offering up the hourly rank tracking tool for free testing.

As much as I love SEO tools – and have gone on record saying this many times – there are a lot out there, busily content marketing on us. Thing is, it’s not always great at keeping the conversation going – and that’s exactly what we wanted to do.

The interest the first image gained and the suggestions which flowed from there were pretty motivating and frankly we want more of it. But I’d be lying if I said it was all altruistic on our part, links & beverages repaid in kind never go a miss.

The Problem With SEOs (sometimes)

The SEO industry is one of the most sharing/open in some respects, people regularly share their research/ideas with the world readily. But, on the flip-side of that, we all have a tendency to believe everything a little too unthinkingly somethings.

In our defence (mine included), validating what we learn from blogs or even what Google tells us is hard work and it leaves us too open in many ways. Here’s a (slightly paraphrased) tweet from Russ Jones which describes our stance behind hourly almost exactly.

Russ Jones' frustrations with the SEO industry via Twitter.

What this doesn’t change is the inescapable – we rely on knowledge share, we just have to all work a little harder and a little more transparently in order to be confident that what we’re sharing is (to the best of our knowledge) legitimate.

I am not saying that hourly is the solution to this problem, but we do hope it can help make things better.

Test #1 – JavaScript Indexing Drag Race

In March I published the findings from one of our earliest tests which looked to see whether or not JavaScript rendered content really took longer to be crawled/indexed than that in the source.

If you want to read more, you can do so here – but below is the condensed version.

The JavaScript indexing drag race experiment too long; didn't read summary.

One crucial element to highlight here is that a few people spotted a potentially significant flaw in the test (which we’d hinted at above). That is; having the content indexed via a ‘fetch & render’ request in GSC was not a great approximation of how it happened “in the wild” and our results were a bit optimistic.

Bummer, right? Wrong –  it was exactly what we wanted.

Research, Hypothesis, Test, Feedback, Re-test - The testing cycle we're following with the hourly tool.

It gives me some more work to do, but it gets us a little closer. We all need to do more of this within SEO.

We’ve got a number of tests bubbling away and some exciting projects on the horizon, but it’s a buzz to see more coming out from the tool.

Other Findings

It isn’t just through the larger tests where we catch glimpses of where this more detailed method of rank tracking can be beneficial, many of the testers we’re working with so far have also shared what they’re seeing too.

What You Miss With Daily

Will shared (via Twitter) this screenshot, which clearly shows that if a daily rank tracker checked at the same time every morning (7am) it would have seen the client ranking on page 8/9, whereas at other points during the day it was as high as page 4/5 – which would have only been seen hourly.

Will O'Hara's findings that daily rank tracking would have missed these ranking fluctuations.

This isn’t the difference between driving traffic and not driving traffic (which would have been far more significant), but anyone reading this data or reporting on it, would behave very differently.

PaddyPower Ranking Troubles

This next example was observed around a suspected update (although, this was an isolated case, so don’t get too excited), what we see is one page ranking constantly and then 3pm in the afternoon what was a steady page one term, was thrown into complete disarray.

PaddyPower ranking problems through hourly

After the highlighted point at least 8 other pages suddenly started competing for the same term – removing any chances of consistent/useful traffic. We have dug into this one in a little more depth (not a client, or affiliation), and these are all lower-quality content pages which have fairly significant internal linking issues.

Monitoring Site Migrations

This one is more in the QI (Quite Interesting) category, but when we watched a HTTP > HTTPS migration process for a particular piece of content we saw the following.

  • Consistent rankings for HTTP version (position 5)
  • 4 hours of fluctuations as the HTTP/S versions competed
  • and then finally the HTTPS ranks outright, 1 position higher – win!

Monitoring a site migration through hourly

What can you do specifically with this information? Not a great deal. Although seeing a problem up-to 23 hours sooner will have its advantages.

Identifying Flux

Something we have been observing recently is that hourly tracking gives a near real-time view of fluctuations across the keyword set*. Some may say it’s noisy, others may say it’s “higher definition” – here’s a comparison of Mozcast 7 days vs Hourly 7 Days:

Mozcast flux charts compared to hourly.

A point to note here, Mozcast is US-based, whereas ours is entirely UK, so the comparison shouldn’t be used for anything meaningful. What’s more fascinating is the period of high-flux on the right of the bottom graph is the broad core update brewing (we think).

*NB it’s ~80 terms now, not a huge sample-size

What Next?

There’s a lot on the horizon for hourly, enough to keep us pretty busy, but feature-wise:

  • Mobile tracking
  • Invite more users onto the platform
  • Universal rankings/more SERP features

For now, if you’re interested click here and register your interest. We can’t guarantee everyone will be let on, but emails/tweets telling me all the cool experiments you have planned may help!

Any other questions, contact myself or Simon via Twitter if you have any questions.

View the Slides

Finally, if you want to view the original slides from Search London (which also contain some stats on the platform) you can view them here

Written By
Chris Green is Head of Marketing Innovation at Footprint Digital, a digital marketing agency based in the East of England.
  • This field is for validation purposes and should be left unchanged.