The 'compare-at' price is (mostly) a fiction

If you've ever stared at a £65 legging with a big red "£85" next to it and felt a small jolt of urgency — that's exactly what the strikethrough is designed to do. The problem is that for most of the products we track, that £85 is not a price the item has actually sold at in any recent memory. It's a reference anchor. A marketing number.

We've been pulling price snapshots for 120 activewear products since early January 2026. What we learned looking across the set is that the "compare-at" number — the strikethrough on the retailer's own site — is almost never the price the product typically sells for. It's close to an aspirational ceiling that the brand posts briefly, then discounts against for the rest of the year.

Here's what that looks like in aggregate.

The median across the full 120-product panel comes out around 14 days per year at compare-at. In other words: the typical "was £85" you see on a brand's own page reflects a price the product held for about two weeks out of 52. The other 50 weeks, the product sold for less — sometimes a lot less.

What that does to your sense of "deal"

Picture two people looking at the same legging on the same Tuesday.

• Person A sees "£65, was £85" on the brand's own site. A 24% discount. Nice.
• Person B looks at a chart of the product's price over the last 90 days. The price has been £65 for 73 of those 90 days. Twelve of the remaining 17 days were at £59. The other five were at £85 — those are the "was" days.

Person A thinks they're catching a deal. Person B knows the "deal" is the default state, and that £59 is the number to wait for.

This is what we mean when we say compare-at is a fiction. It's not that the number is fabricated — it's that it represents such a small slice of the product's actual selling life that treating it as the baseline warps every subsequent "discount."

How we compute our own baseline

We throw the compare-at number out entirely. For each product we maintain a rolling 90-day window of observed prices. The relevant statistics for us are:

• P5 — the 5th-percentile price over the window. Roughly "what the product sells for when it's actually on sale."
• P50 — the median. The "normal" everyday price.
• Current — today's price.

A deal, to us, means today's price is within about 5% of P5, and today is meaningfully lower than P50. We don't care what the brand's strikethrough says; we care whether today is genuinely unusual versus the product's own recent history.

This is why the "Lowest in 90 days" badge you see on Gearfinch is a stronger claim than a retailer's sale badge. The retailer is comparing today to an arbitrary anchor. We're comparing today to what the product has actually been selling for.

Brands that play it straighter

Not every brand uses compare-at theatrically. Of the seven we track, two — NVGTN and Born Primitive — use strikethroughs sparingly, mostly during genuine seasonal clears. Alo Yoga is the most aggressive: the Airbrush 7/8 in particular has spent roughly 90% of the last four months below its "retail" number, which makes the retail number functionally meaningless.

We don't assign moral weight to this. It's a pricing choice. But it does mean that if you're trying to figure out whether a specific product is actually cheap today, the brand's own page is the wrong place to ask.

What this looks like in the product

If you open any product page on Gearfinch, the "lowest in 90 days" line is computed against our own history, not the retailer's compare-at. When we say a price is the lowest we've ever recorded for a product, that's literally true to the minute the last scrape ran. It's boring infrastructure, but it's the whole point of the site.

The compare-at number has its uses — occasionally a product really is worth the full retail figure, and seeing how far today's price is from that ceiling is interesting context. But it should never be the number you're evaluating a sale against.

That's what we're here for.