We built repricing for Etsy. It will never raise your price.

Everlyst’s Pricing Intelligence page started as an advisory tool. It showed which listings had gone quiet, compared their prices with similar listings inside your own shop, and left the decision and the edit to you.

Now it can apply a reviewed percentage reduction directly, to one listing or as many as 250 in a request.

There is one thing that action will not do, at any percentage or under any configuration: raise a price. Pricing Intelligence may still point out that a strong seller priced below its comparison average has room for review, but Quick Apply only executes reductions. That boundary is a design decision, not a missing control.

What repricing usually means

Repricing as a category comes from large retail marketplaces, where several sellers may offer the same fungible item. A competitor-price feed changes, a rule reacts, and a small price difference can affect which offer wins the sale. At that scale, continuously watching thousands of products is a job for software.

On Etsy, dynamic pricing products adapt that idea to a different marketplace. Rules respond to a seller’s own sales performance, moving a price up or down while controls such as price floors limit the result.

Evlista is one established example. Its dynamic pricing responds to sales performance, respects price floors, and requires seller approval before a change is applied. If bidirectional, rules-based repricing is central to your workflow, Evlista offers it. Everlyst does not.

Our difference is not about whether that approach can be implemented carefully. It is about which pricing decision software should make easy on Etsy.

Why we kept the write path one-directional

Etsy catalogs rarely provide the clean competitor set that retail repricing assumes. Handmade, personalized, vintage, and small-batch listings differ in materials, positioning, presentation, and fulfillment. Everlyst can compare a listing with the other active listings in its shop section, but that average is context, not a claim that the products are interchangeable.

Sales history is noisy too. Many listings do not sell often enough to reveal a dependable demand curve. Seasonality, promotions, traffic shifts, and a handful of orders can all make one month look meaningfully different from another. A seller may reasonably decide that strong demand justifies a higher price, but a rule cannot know the seller’s costs, capacity, positioning, or tolerance for risk.

There is also a search consideration. Etsy says listing engagement, including how well a listing converts, contributes to placement in search results. A price increase can be the right business decision, but if it weakens conversion, the effect may extend beyond margin on the next order. That is a tradeoff a seller should evaluate, not a background action Everlyst should take for them.

So Pricing Intelligence can surface evidence for a possible increase, but the direct action on the page has a narrower job: help a seller clear stale, relatively expensive listings after review.

When a reduction recommendation appears

Crossing the stale threshold is not enough by itself. Everlyst recommends a reduction only when all of the following are true:

• The listing has gone without a sale for the number of days you configured.
• Its current price is at least 15% above the applicable comparison average.
• There are enough active, priced listings to support that comparison.

The comparison uses the listing’s shop section when that section has a sufficient cohort. Otherwise it falls back to the shop-wide average. If the shop itself does not have enough comparable listings, Everlyst suppresses the recommendation instead of manufacturing certainty from too little data.

You can define “stale” as anywhere from 7 to 365 days. A listing that has never sold can qualify too, using its creation date in place of a last-sale date, but it still has to meet the price and comparison requirements.

When you are ready to act, select individual listings or use Select recommended to gather the reduction candidates. Choose a discount from 1% to 99%, review the result, and confirm. Nothing runs on a schedule and nothing applies without that confirmation.

The rails around every reduction

The mechanics matter more than the label.

You pick the percentage. Everlyst does not snap a listing to its section average or treat that average as a target. It is evidence for your decision, not an instruction to the write path.

Variation price spreads survive. Each offering is reduced individually. If one option starts at $18 and another at $34, both move by the selected percentage instead of collapsing to one flat price.

You can set a floor and a price ending. The floor prevents an adjusted offering from falling below the minimum you choose. Optional rounding supports .99 and .95 endings. Because rounding or a high floor could otherwise push the result back above its starting point, the calculation caps every offering at its original price. Quick Apply never raises one.

Every change is confirmed and queued. The dialog previews the representative before-and-after price and states the floor and rounding rules before the job starts. The adjustment then uses Everlyst’s tracked bulk-operation queue rather than writing silently in the background.

A snapshot is captured first. Each changed listing gets a pre-adjustment inventory snapshot. The operation appears in Backup & Restore, and restoring it changes the retained inventory prices back without replacing unrelated listing fields.

Recently changed prices cool down. A price change made through Everlyst records when it happened. Pricing Intelligence then holds that listing out of reduction recommendations for half of the configured stale window. Set the window to 120 days and a recently adjusted listing waits 60 days before the page considers another recommendation.

Deterministic, reviewed, and reversible

The recommendation is arithmetic on your own catalog and sales history. The same inputs produce the same result: a known stale threshold, a known comparison cohort, and a known 15% price gap. You can inspect every part of that reasoning on the page.

There is no model setting your prices and no autonomous rule engine waiting for a sales signal. The page can tell you that a price looks high or low relative to your own catalog. Only you can decide whether that evidence fits the listing, and Quick Apply only carries out the reviewed reduction you specify.

That makes the feature narrower than bidirectional dynamic pricing on purpose. It finds stale listings whose prices stand well above a defensible internal comparison, lets you reduce them in one reviewed pass, preserves the variation pricing you built, and keeps the operation reversible.

Pricing Intelligence is available on Pro.