Competitor price monitoring: match the right product, trust the feed

Competitor price monitoring means automatically collecting the prices, stock status and promotions your rivals publish, tying each of those listings to the right product in your own catalogue, and telling you when something changes. The collecting is the easy half. The two things that decide whether it works day to day are matching the right product and knowing the moment a source stops returning true data.

What competitor price monitoring actually tracks

A price on its own is thin. A pricing decision needs the price a shopper would pay today, at that seller, for that exact product, and whether they could buy it at all.

• Shelf price. The price after on-page discounts, not the crossed-out list price sitting next to it.
• List price and discount depth. A competitor at $179 against a $229 list price is running a 21.8 percent markdown. A competitor at $179 against a $179 list price is not on promotion at all, and those two situations call for different responses.
• Stock status. In stock, out of stock, backorder, or limited availability. A price nobody can buy is not a price you need to beat.
• Promotions and coupon text. Percent-off banners, bundle offers, and threshold offers such as free shipping over $50.
• Shipping cost and thresholds, because the delivered price is what competes, not the shelf price.
• Seller identity on marketplaces. A marketplace listing price belongs to whichever seller is currently winning the offer, and that seller can change several times a day.
• Variant-level price. Size, color, capacity and pack count each carry their own price, and averaging across them hides the one that is losing.

Collection covers storefronts and marketplaces together, so an authorized dealer and a third-party marketplace seller appear in the same view. The price monitoring module stores every observation with a timestamp and its source URL, which is what makes an argument about a number resolvable later.

Product matching: the step that decides whether any of it is true

Almost every wrong number in a price feed traces back to a wrong match. The scraper worked. It just read the price of a different product. Matching runs as a waterfall, strongest signal first, and every result carries a confidence score.

1. 1

   Identifier join

   GTIN, UPC, EAN and MPN are matched exactly. Where a retailer publishes an identifier there is nothing to interpret, which is why branded goods are easier to monitor than private label.

2. 2

   Title and attribute embeddings

   Where no identifier is published, the product title and its structured attributes are compared as embeddings, so that "18V brushless drill kit, 2 x 4.0Ah" and "Brushless Drill Kit 18 volt with two 4Ah batteries" resolve to the same product instead of two.

3. 3

   Image comparison

   For categories where titles are unreliable, such as furniture, apparel and home goods, product imagery confirms or rejects a candidate the text was unsure about.

Corrections are not thrown away. When you reject a match or point at the right listing, that decision feeds back into matching, so the same mistake is not offered again next week. The product matching guide covers the harder cases: bundles, multipacks and refurbished listings.

Refresh rate: how often price tracking software should check

Match the refresh rate to how fast your category moves and how fast you can act on what you see. Collecting four times a day and approving price changes once a week wastes both. Refresh is daily on Starter, twice daily on Growth, four times daily on Scale, and hourly on Enterprise.

Here is the arithmetic that settles it. A competitor runs a six-hour flash sale. A once-daily check catches it only if the check lands inside those six hours: 6 hours out of 24, a 25 percent chance. A four-times-daily check sees it nearly every time. In categories with frequent short promotions, daily collection is sampling, not monitoring.

• Daily is enough for furniture, industrial supply, building products and most private-label assortments, where prices move on a weekly or seasonal cadence.
• Twice daily or faster for consumer electronics, small appliances, and anything with heavy marketplace competition, where the winning offer changes within the day.
• Hourly where you reprice automatically and a stale number would push a live price in the wrong direction.
• Refresh rate and product count are the only two things that change between plans. No capability is held back for a higher tier.

Four ways to collect competitor prices, compared

Most teams reach monitoring software after trying two of these. The difference that matters is not features. It is how each approach fails, and who notices.

How a competitor price tracker breaks, and how this one heals

Price feeds rarely fail loudly. They fail quietly, and a number keeps arriving on schedule. That is the whole problem. Here are the four failures in mechanical detail, because a vendor who cannot describe them has probably not solved them.

The four ways a feed goes wrong

• The price moves into a client-side component. The server returns a page where the price element is empty and populated later by JavaScript. A naive extractor stores null, or repeats yesterday's value, and the product appears frozen at a price that expired a month ago.
• The selector drifts onto the wrong node. A layout change puts the strikethrough list price where the shelf price used to sit. The feed reports $249 for a product selling at $179, and a rule that positions you under the market average prices you out of contention while the dashboard looks healthy.
• The product URL changes. A retailer moves /products/oak-patio-chair to /p/48812. The old URL returns a 404, that competitor silently drops off the product, and the market average reweights toward whoever is left.
• Bot mitigation returns a 200. A challenge or interstitial page comes back with a success status code. The extractor finds no price, or a price of $0.00, and plausibility validation is the only thing between that value and a live repricing rule.

What happens instead

• schema.org Product structured data is read first wherever a site publishes it, so the price comes from the same feed the retailer hands to search engines rather than from a fragile CSS path.
• A per-domain extraction recipe is the fallback, and when a site's layout changes that recipe is regenerated automatically instead of waiting in a ticket queue.
• When a product URL moves, the match is re-discovered by identifier rather than by URL, so coverage survives a site migration.
• Every price passes plausibility validation before it is stored. A zero, a shipping fee read as a price, or a fortyfold overnight jump is quarantined rather than published.
• A coverage and freshness dashboard shows, per source, when it last returned a good price and what share of your catalogue it currently covers.
• You are alerted when a source degrades, before you would have noticed it yourself.

Alerts that reach a human at the right moment

An alert is useful only if it contains enough to decide without opening anything else. Every alert carries the product and your price, the competitor and their new price, the move in dollars and percent, whether that competitor is in stock, the resulting gap, and a link to the source page with its timestamp.

The harder half is thresholds. Alert on boundaries you would act on, not on every cent, or the channel becomes noise inside two weeks. A starting set for most catalogues:

1. 1Any in-stock competitor crossing your price, in either direction.
2. 2A move larger than a set percentage on a monitored product. Three percent is a sensible starting point for general merchandise, one percent for high-velocity electronics.
3. 3A monitored product where every in-stock competitor is now below you, which is a different signal from one competitor going low.
4. 4A source that has not returned a good price for two consecutive refresh cycles.
5. 5Any listing below MAP, immediately, with screenshot evidence attached. That workflow lives in MAP monitoring.

Alerts go to email and Slack, and a weekly digest summarizes what moved for people who do not want a live feed. Merchandising usually reads the digest.

What to do with the data once you trust it

Collection is a means. Three things are worth doing with a trustworthy feed, roughly in this order.

Measure position before changing anything

Price index expresses your price against the market average, where 100 means at market. Position mix shows what share of your assortment sits lowest, matched or above. Run both for two weeks before touching a price. Teams routinely find the products they worried about were fine while a quiet corner of the catalogue was thirty points high. The price index and positioning guide explains the calculation.

Reprice with guardrails, not with hope

Rules carry hard floors: margin floor, cost floor, MAP floor and a maximum daily change. Output goes where you trust it to go, which for a new account means a review queue, then CSV export, then a direct push once the numbers have earned it. Every change is kept in full history, so a price can be explained months later. See repricing software for the rule builder.

Watch the pattern, not the point

A single low price is an event. A price that drops every second Thursday is a cadence. Reacting to the trough of someone else's promotional calendar costs you margin twice a month and teaches them nothing. A quarter of price history makes the difference obvious.

How to choose competitor price monitoring software

Most vendors show a similar screenshot of a grid with green and red numbers. These questions separate them, and each has an answer you can check inside a trial rather than take on trust.

1. 1Show me a product with a wrong match, and the workflow to fix it in under a minute.
2. 2How do I discover, without asking you, that a source stopped working last Tuesday?
3. 3What is read first on a page: publisher structured data, or a CSS selector someone wrote by hand?
4. 4What happens to a price of $0.00 between collection and my repricing rule?
5. 5Is stock status collected, and can out-of-stock competitors be excluded from the market average?
6. 6Which capabilities are gated behind a higher plan, and which are included everywhere?
7. 7Can I export raw observations with timestamps and source URLs, or only summary charts?
8. 8How long from signup to the first matched competitor price on a real product of mine?

Our answer to the sixth question is worth stating plainly: capability is never the upsell. Matching, MAP monitoring with screenshot evidence, repricing, analytics, API and webhooks, every integration and unlimited users are on all four plans. You pay for products monitored and refresh rate.