How eBay’s Pricing Algorithm Actually Responds to Repricing

You dropped your price by 10% expecting more views. Instead, views went down. What happened?

eBay’s search algorithm — Cassini — doesn’t work the way most sellers think. It’s not a simple “lower price = higher ranking” equation. Understanding how Cassini actually processes price changes is the difference between strategic repricing and shooting yourself in the foot.

Cassini 101: What It Optimizes For

Cassini’s job is to show buyers the listings they’re most likely to purchase. Not the cheapest listings. Not the newest. The most likely to convert.

That conversion probability is based on a cocktail of signals:

Relevance: Does your title and item specifics match what the buyer searched for? Seller performance: Your shipping speed, feedback score, defect rate, and return handling all feed into this. Listing quality: Number of photos, completed item specifics, detailed description. Conversion history: How often does this listing get clicked? How often do clicks result in purchases? Recency: Newer listings get a temporary boost. Older listings need to prove their value through conversion signals.

Price is a factor — but it’s filtered through all these other signals. A well-optimized listing at a higher price can outrank a poorly-optimized listing at a lower price.

What Happens When You Lower Price

Lowering price triggers a few things in Cassini:

Short-term visibility bump: eBay temporary boosts repriced listings, similar to the new-listing boost. This is the “repricing works!” feeling sellers get.

Watcher notifications: If your listing has watchers, they get notified of the price drop. If they buy: great, positive conversion signal. If they don’t buy (which is common): that’s a bunch of impressions with no conversion. Cassini notices.

Competitive repositioning: If you’re now the lowest-priced comparable listing, you may rank higher in “Price + Shipping: lowest first” sorts. But most buyers use “Best Match,” which is Cassini.

Margin impression: Repeated price drops can signal desperation to the algorithm. A listing that started at $50, dropped to $45, then $40, then $35 shows a pattern of non-conversion at multiple price points. That’s a negative signal.

The takeaway: one strategic price drop can help. Serial price drops without other improvements usually don’t.

What Happens When You Raise Price

Counter-intuitively, raising your price doesn’t always tank your visibility. Here’s why:

If your listing has strong conversion signals at the current price — good click-through rate, watchers, questions from buyers — and you raise the price modestly (5-10%), Cassini doesn’t punish you because your conversion signals are still positive.

The risk is raising price on a listing that already has weak conversion. A listing with few views that gets more expensive will just get fewer views. The algorithm doesn’t reward optimism.

When raising price works: You have comparable listings selling at higher prices. Your listing is demonstrably underpriced relative to completed sold listings. And your listing quality (photos, specifics, feedback) supports the higher price.

Repricing Frequency and Its Effect on Ranking

How often you reprice matters:

Too frequent (daily or every few days): Creates algorithmic noise. Cassini can’t establish stable conversion signals because the variables keep changing. Each change resets the evaluation period.

Too infrequent (never): Your listing ages, loses its recency boost, and potentially falls behind market changes.

Right cadence (every 7-14 days, if repricing at all): Let each price point run long enough for Cassini to evaluate its performance. If views and conversion are acceptable, hold. If they’ve degraded, reprice once with a meaningful adjustment (not $0.50 — think 10-15%).

The worst thing you can do is lower price by $1 every three days. It doesn’t create enough impact to improve conversion, but it does create a “declining” pattern that Cassini doesn’t like.

The “Revise vs End-and-Relist” Question

When a price change isn’t enough, should you revise the listing or end it and relist entirely?

Revise keeps your existing listing: same item number, same watchers, same conversion history. Good for healthy listings that need a price adjustment.

End and relist creates a new listing: new item number, fresh recency boost, but you lose watchers and accumulated conversion data. Good for listings that have been stale for 60+ days with no engagement.

The watchers question is the tiebreaker. If your listing has 10+ watchers, revising preserves that social proof. If it has 0-2 watchers after 60 days, there’s nothing to preserve — relist.

Data-Driven Repricing vs Gut Feel

Most sellers reprice based on gut feel: “Hasn’t sold yet, must be too expensive.” But without data, you don’t know if price is the problem.

Your listing might have low views because of a bad title (not a pricing issue). Or it might have high views but low conversion because your photos don’t inspire confidence. Or it might have decent engagement but just needs more time in a seasonal category.

Before repricing, ask:

• Are my impressions and views consistent with other listings in this category?
• Am I getting clicks (impressions → views) at a normal rate?
• Am I getting purchases (views → sales) at a normal rate?

If impressions are low, your title and item specifics need work. If clicks are low, your main photo needs improvement. If clicks are fine but no sales, then yes — price might be the issue.

Building a Repricing Cadence

Here’s the system I use:

Week 1-2 after listing: Don’t touch it. Let Cassini evaluate the listing naturally. Watch views and watchers.

Week 3-4: First assessment. Are views in line with expectations for this item and category? If not, improve title/photos/specifics first, not price.

Week 5-8: If views are healthy but no sale, consider a 10-15% price reduction. One adjustment, let it run for another 2 weeks.

Week 9+: Decision time. Either this listing needs a significant price drop (20%+), a full relist with improved content, or acceptance that it’s a long-tail item.

Tracking this cadence for each listing manually is tough at scale. What you want is a system that flags listings by age and engagement level, so you know which ones need attention without checking each one individually.

That’s the difference between reactive repricing (reacting to slow sales) and proactive pricing management (systematically optimizing across your entire inventory).