Does Google Watch How Long Patients Stay on Your Site?
You don’t grade a resident’s chart note by how long it took to write.
A three-line assessment that nails the diagnosis beats a rambling page that circles the point, and every physician learns that early.
Length and duration measure effort, which is a different thing altogether from whether the encounter did what it needed to do.
Plenty of practices run their websites on the opposite assumption. Someone tells the doctor that Google rewards pages patients linger on, so the fix becomes more paragraphs, another embedded video, an FAQ split across three extra clicks so people stick around longer.
And the goal quietly shifts. Answering the patient’s question gives way to holding their attention, which is backwards, and it’s worth being precise about why.
What Google has actually said about this
“Dwell time,” meaning how long someone stays on a page before clicking back to the search results, has circulated as an SEO concept for years. It’s intuitive. A patient who reads for two minutes seems more satisfied than one who bounces in five seconds.
The trouble is that Google has never confirmed it as a ranking signal, and its engineers have said the opposite when asked directly.
In a Reddit AMA, Google’s Gary Illyes described the actual ranking system as far simpler than the dwell-time folklore suggests, built around relevance signals rather than behavioral proxies like time-on-page.
John Mueller has been blunter still. He’s stated plainly that dwell time isn’t something Google’s ranking systems use.
There’s also a practical reason it couldn’t work reliably even if Google wanted it to.
A patient who finds your hours in four seconds and calls your office has succeeded completely. A patient who spends three minutes on your page because the appointment-request button is broken has failed, and took longer doing it.
Both produce a number, long or short, and the number by itself can’t tell you which one happened.
Google’s own people have pointed at exactly this measurement problem. Time on page conflates satisfaction with confusion, and a search engine that ranked pages by it would be optimizing for the wrong thing on purpose.
What the engagement dashboards are actually selling
Plenty of marketing vendors sell scroll-depth widgets, session-replay tools, and “engagement scores” that promise to lift your dwell time and, by extension, your rankings. The engagement number is real, but the rankings claim attached to it doesn’t hold up.
Watching a screen recording of a real patient’s mouse movements is genuinely useful. It just isn’t an SEO tactic, and treating it like one leads a practice to redesign pages around the wrong target.
The tell is in how the pitch gets framed. If a report shows “average time on page up 40 percent” as evidence the SEO is working, ask what you’d ask of any surrogate marker.
Up compared to what outcome?
A longer visit that ends in a phone call is good news. A longer visit that ends with the patient giving up because they couldn’t find your address is the same number pointing the other way.
The dashboard can’t tell those two apart, and neither can a page score built out of it.
What time-on-page is genuinely useful for
The metric’s fine. It’s just been pointed at the wrong job for years. Time-on-page and its cousins (scroll depth, rage clicks, exit rate) work as diagnostics, and what they turn up is the pages that are failing patients.
Used that way, the pattern is easy to read:
- A page with very short visits and a high bounce rate often means the content answered the search query right there in the preview snippet and there was nothing left to read. Or the page loaded slowly and the patient left before it rendered. Both look identical in the raw number, so you have to go check which one it is.
- A page with long visits and no calls or form submissions usually means a patient is stuck, reading through dense text hunting for something the page never states plainly, like whether you accept a particular insurance plan or perform a particular procedure.
- A page where visitors scroll straight to the bottom and leave often means the thing they came for – office hours, a phone number, a specific condition – was buried under several paragraphs of general practice description.
None of that moves a ranking, though every one of those fixes changes whether the next patient who lands on the page picks up the phone.
Pull your five busiest pages and look for the stall
Open whatever analytics your site already has (Google Analytics, or a free session-recording tool) and pull up your five highest-traffic pages. Usually that’s the homepage, a couple of condition or service pages, and your contact page.
Look for the ones where visitors stay a long time and still don’t call or fill out the appointment form. Those are your shortlist. On each of them a real patient is reading and reading and still not finding the thing that would make them reach for the phone.
Fix what’s missing there. The ranking sorts itself out, and the patient sitting in your waiting room next month was always the metric worth chasing.
Questions practices ask about this
My SEO report shows average time on page is up. Does that mean the work is paying off?
Not on its own. Time on page moves for good reasons and bad ones, so ask for call and form-submission counts on the same pages over the same period, printed right beside it. If your vendor can't connect the two, that number is decoration. A useful report names which pages changed, what changed on them, and what patients did next.
How do I know if my practice has this problem?
Pull your five busiest pages and compare how long visitors stay against how many calls or form submissions each one produces. A page that holds people for two or three minutes and still generates almost nothing is where patients are stalling. Session recordings in a tool like Microsoft Clarity show you the exact spot they give up. Check your contact page first.
Can my office manager do this, or do I need to hire someone?
Your office manager can handle the first pass in about an hour. Pulling your five busiest pages and noting which ones hold visitors without producing calls takes no technical skill, just access to your analytics. The fix is the harder half, since rewriting a service page or moving a phone number above the fold means editing the site itself. Hire out only that second half if you can't edit the site yourself.
How long after I fix a page should I be able to tell whether it worked?
Give it four to six weeks on a page with steady traffic, and longer if that page only gets a handful of visits a week. You're comparing calls and form submissions before and after, so each side needs enough sessions for the difference to mean anything. A quiet page can take a couple of months before the numbers settle. Write down the date you made the change.
Photograph: RDNE Stock project / Pexels