Both numbers are real, and Pinterest outbound clicks were never going to line up with Google Analytics sessions. Pinterest records an outbound click at its own redirect, the moment the tap registers, before your server has been asked for anything at all. GA4 records a session only after your page has loaded and your tag has run. Everything that can go wrong in between comes out of the second number and never out of the first. So 1,800 against 900 feels like theft and is closer to arithmetic. Pinterest does not charge you for organic outbound clicks, which means nobody gains anything by inflating them.
The reasonable next thought is that one of the two is lying, and obviously it is the one with something to gain. You were expecting me to agree, right? Neither of them is lying. There are four separate leaks between the tap and the session, every one of them is measurable, and by the end of this you will be able to put a number on each for your own site instead of arguing with a dashboard.
Table of Contents
The two numbers are counting different moments
Pinterest defines an outbound click as the number of times people perform actions that lead them to a destination off Pinterest. Read that definition again slowly. The thing being counted is the action rather than the arrival. The event is logged at Pinterest’s own redirect, which fires the instant a finger lands, and the counter has already moved before your DNS has even been resolved.
GA4 works from the other end entirely. A session begins when somebody views a page while no session is already running, and Google records a session_start event at that point. The default session timeout is thirty minutes of inactivity. Nothing in that definition can happen until your HTML has been delivered and your GA4 tag has executed, so anything that interrupts the journey deletes a session while leaving Pinterest’s number exactly where it was.
Pinterest adds a wrinkle that catches people doing month on month comparisons. Outbound clicks for a pin are measured starting from the date the destination link was added, which is not always the date the pin itself was created. Edit a link on an old pin and its click history restarts from that edit, so your date ranges stop lining up with anything.
Referrer stripping files real Pinterest visits under Direct
This is the biggest single cause, and it is the one that makes people angriest, because the traffic genuinely did arrive. It just arrived without a name tag on it. Pinterest’s in app browser and its mobile redirects frequently drop the referrer header, and GA4 classifies a hit with no referrer as Direct, since its rule for Direct is a source of (direct) with a medium of (none) or (not set). Organic Social only catches a visit when the source matches a known social domain or the medium says social. No referrer means no match. A genuine Pinterest visitor lands in the wrong bucket, your Social number reads low, and your Direct number quietly grows to match.
You can watch this happening in your own property. Open your Direct traffic and look at the landing pages. Direct is supposed to mean bookmarks and typed addresses, which in practice means your homepage and a few pages people know by name. If Direct is landing on a deep article URL with a forty character slug that no human being would ever type from memory, that is misfiled referral traffic. If that same URL is one you pinned, you have found your missing Pinterest visitors.
Consent banners and blockers take a share that depends on geography
If your reader declines analytics cookies, or runs a browser that blocks the GA4 endpoint by default, the visit still happens and the session never appears. The size of this loss has nothing to do with Pinterest and everything to do with where your audience lives. A site read mostly in Germany or France, where consent banners are enforced hard and refusal is common, will show a far wider gap than a site read mostly in the United States. Two identical Pinterest strategies produce two very different looking dashboards for that reason alone.
Measure your own rather than borrowing somebody else’s benchmark. Look at the country breakdown of your audience inside Pinterest, then look at the country breakdown of your sessions in GA4. If Europe is a large slice of the first and a small slice of the second, consent is eating your numbers before Pinterest has anything to do with it.
GA4 folds several taps by one person into one session
Pinterest traffic produces exactly the behaviour that the session definition collapses. Somebody is scrolling a board, taps into your recipe, backs out, taps your second pin, backs out again, taps a third. Pinterest counted three outbound clicks, because three actions happened. GA4 counted one session, because all three arrivals fell inside the thirty minute window and no session ever ended in between. A saver in a browsing mood can generate five or six clicks against a single session, and a board stuffed with your pins makes that the normal case rather than the odd one.
You can size this leak directly. Compare Users against Sessions for your Pinterest traffic. If sessions per user sits well above one, you have a repeat tapping population, and every extra tap in that ratio is an outbound click that was never going to get a session of its own.
Some taps never became a page view at all
The last leak is the boring one. A person taps, the page starts loading over a slow mobile connection inside an in app browser, and they swipe away before your tag fires. Old pins pointing at URLs you have since changed produce a redirect chain or a 404 instead of a session. A consent wall that blocks rendering until a choice is made loses everyone who makes no choice. None of this is Pinterest’s doing, and all of it subtracts from the same side of the comparison.
Server logs settle this one for good. Count requests to the URL across the same window and compare them against GA4 sessions for that URL. Requests that arrived and never became sessions are your blocked and abandoned group, sitting there in plain text.
Size your own gap in an afternoon
- Add UTM parameters to ten pins pointing at pages you can watch, with a source of pinterest and a medium of social, so GA4 stops having to guess.
- Wait a week, then pull outbound clicks for those ten pins from Pinterest analytics.
- Pull sessions for that campaign in GA4 across the same dates, and widen the window by a day at each end so that timezone differences between the two tools cannot distort the comparison.
- Pull Direct sessions landing on those exact URLs, then compare against Direct for the same URLs in the month before you pinned them. The increase is your referrer loss.
- Check sessions per user for that campaign. Anything above one is the folding effect showing up.
- Pull request counts for those URLs from your server logs. Requests minus GA4 sessions is consent, blocking and abandonment added together.
Add the misfiled Direct back onto Social, allow for the folding ratio, and the 900 you started with usually explains most of the 1,800 without anybody having lied to you. Whatever is left over is your genuine loss, and it will be a great deal smaller than the number that made you suspicious in the first place.
Decide which number to run your business on
Here is where getting it wrong costs real money. Use GA4, or better still a server side count, for anything you are deciding about your own site, because those numbers describe people who genuinely reached your page. Use Pinterest’s outbound clicks for comparing pins against each other, since the measurement error is roughly consistent across pins on one account. Never add the two together. Never hand a client a Pinterest outbound click figure and call it a visit, because when they check it against their own analytics you will spend the meeting defending a number you cannot defend.
And do not close the channel over this. Plenty of people have seen a gap like yours and walked away from Pinterest, which turned traffic that was arriving under the wrong label into no traffic at all. Your gap is a measurement problem with four named causes and a method for pricing each of them. Go and measure it before you decide anything.
