Marketing tools are damaging your SEO. And how to fix it.

It's no secret that a site's “performance” is a ranking factor for search. But how does Google actually measure it?

Well, open up Chrome DevTools, click the Audits tab, then Run Audits and you'll see that “Performance” is made up from 6 metrics:

And if you dig into Chrome's GitHub page you'll see that they've all got different weights attached:

Most casuals assume that performance is all about how fast a site displays on the screen — “First Meaningful Paint”. But if you look at the weights you'll see that “Time to Interactive” — how long it takes for a site to become interactive — is far more important.

Essentially, Google prioritises “interactivity” over “visibility”. Even if your page appears fast, 3rd party scripts loading in the background (delaying user interaction) will see you penalised.

ToDesktop

One example of this is ToDesktop. Their site content loads in < 1s, but run it through a Google Audit and its “Time to Interactive” is a whopping 8.8 seconds.

The culprit (as is often the case) is 3rd party scripts from shiny marketing tools: Intercom for chat, Segment for data, Hotjar for heatmaps, etc ...

Solving the problem

So, what can you can do about this? Well, ToDesktop's founder, Dave, came up with a clever solution. First, load the page without any heavy scripts. Then wait for a scroll event, then wait a further second, and then load in Intercom, Segment, etc ...

The before and after results are quite revealing:

“Time to Interactive” plummets from 8.8s to 0.9s — mission accomplished. Interestingly, “Paint” numbers remain exactly the same. The casual observer may notice no real difference in performance. But Google certainly does.

The code snippet is pretty straightforward:

Does this stuff make a difference?

Well, Dave made this change on June 14th. Since then ToDesktop's organic clicks and impressions have nearly tripled and Dave tells me that he's jumped from 25th to 11th for the search “website to desktop app”.

It's impossible to isolate how much of this can be attributed to improved “performance”. But there's certainly a clear correlation.