We Measured 2,871 Pinterest Pins: What Actually Gets Saved

Most Pinterest advice is repeated, not measured. We had an unusual chance to check some of it: URL2Pin generates pins for other people’s products, publishes them through the official Pinterest API, and syncs the results back. That leaves a dataset of pins that were written by the same system, for different products, by different people — which makes the copy differences easier to isolate.

This post reports what we found across 2,871 published pins carrying roughly 103,000 impressions. Everything below is stated as a multiple rather than a raw rate, for reasons we explain in the limitations at the end.

1. Titles starting with a number were saved 2.9× more

This was the largest effect we measured, and the samples are close enough in size to trust.

The position of the number mattered. “7 Ways to Style a Narrow Hallway” outperformed “How to Style a Narrow Hallway in 7 Ways”, even though both promise the same thing. A number at the front sets an expectation before the reader has finished scanning.

The question row is the one to treat carefully. Across 111 pins and roughly 2,000 impressions we recorded no saves at all — but at the rates seen elsewhere in the dataset you would only expect two or three, so “none” is suggestive rather than proven. We report it because it points the same way as the rest of the table, not because 111 pins settles it.

2. Longer titles beat short ones

This one runs against the usual guidance to keep pin titles punchy.

A plausible reading: a longer title has room for the specific detail that makes a pin worth saving — the room, the season, the constraint, the number. “Small bathroom storage” gives a browser nothing to file it under. “9 small bathroom storage ideas for a rental you can’t drill into” does.

If you want to test title variants without writing each one by hand, the free pin title and description generator produces several angles from one URL.

3. Etsy products were saved 2.8× more than Amazon — and clicked 3.6× less

The most interesting result was not about copy at all. Splitting by where the product came from produced an almost perfect inversion between saves and outbound clicks.

People collect handmade and buy commodity. A ceramic mug from a small shop gets filed onto a board for a kitchen that does not exist yet. A phone charger gets clicked or ignored. Neither behaviour is better — they are different funnels, and they reward different pins.

We pulled this one apart separately, because it changes what you should optimise for depending on what you sell: why Etsy products get saved more than Amazon products.

4. Tuesday and Saturday outperformed the rest

Pins published on a Tuesday or a Saturday were saved about 2.6× as often as pins published on the other five days, across roughly 400 pins per weekday.

We would treat this as the weakest finding here. Posting day is entangled with what people happen to post — someone scheduling a seasonal gift guide on a Saturday is not running the same experiment as someone posting a kitchen gadget on a Thursday. The effect is large enough to mention and not clean enough to build a schedule on.

What we would actually change

• Lead titles with a number wherever the number is honest. Do not invent one to fit a format.
• Aim for roughly 55–69 characters rather than trimming to a slogan.
• Put the specific detail in — the room, the constraint, the season, the price band.
• If you sell handmade, optimise for saves. If you sell commodity, optimise for the click.
• Do not phrase the pin title as a question. Save the question for the image.

Method and limitations

Data. 2,871 pins published through the official Pinterest API between roughly May and August 2026, carrying about 103,000 impressions in total. Pins with no recorded impressions are excluded, because a save rate needs a denominator.

Why multiples, not rates. Absolute save rates vary enormously by account age, follower count, board topic and niche. A rate from our dataset would not transfer to yours, and would invite comparison against numbers that were measured differently. The ratios travel better.

What this is not. These pins came from one tool, so the writing shares a house style. They skew toward affiliate and product content rather than recipes or weddings. And every result here is a correlation — we did not run a controlled test where the same product got two titles at random. Someone should, and we would happily fund the coffee.

Related reading: impressions but no clicks, the Pinterest troubleshooting hub, or generate a batch of pins from one product URL with URL2Pin.

Using these figures in your own writing is welcome — a link back to this page is all we ask.