Every price monitoring project starts with somebody saying "let's track everyone". It is a reasonable instinct and it is the main reason these projects get cancelled at renewal. A list that is ten times too long costs ten times too much, takes ten times as long to match, and produces a report with so many rows that the person who asked for it stops opening it.
The goal of this lesson is a defensible list: a set of competitors and a set of your own products where you can explain, line by line, why each one is there.
Ranking competitors
Score each candidate on two axes. You want the ones that are high on both.
- 1
Overlap: how much of your range do they actually carry?
Take your top 100 products by revenue and check, by hand, how many of them the competitor stocks. Twenty minutes of clicking gives you a number. A competitor carrying 8 of your top 100 is not a competitor for pricing purposes, however loudly sales talks about them. A competitor carrying 70 of them sets your market price whether you like it or not.
- 2
Threat: does a shopper comparing you to them actually switch?
Price only matters where the alternative is credible. A rival with worse delivery, no stock, or a market the customer will not buy from can be cheaper than you all year and cost you nothing. The honest test is whether your own sales team mentions them in lost-deal notes. If they never come up, they are a benchmark, not a threat — and benchmarks can be checked monthly instead of daily.
- 3
Separate marketplaces from retailers
A marketplace listing is not one competitor, it is a rotating cast of sellers behind one URL, and the price you capture is whoever won the buy box at the moment you looked. That is still useful, but treat it as a market signal rather than as a named rival, and do not build a repricing rule that chases it directly.
- 4
Cut to between three and eight
Almost every retailer we have worked with ends up with a daily list in that range, no matter how big they are, plus a longer tail checked weekly or monthly. If your daily list has twenty names on it, two things are true: you cannot name what you would do differently based on fifteen of them, and you are paying for all twenty.
Ranking your own products
The competitor list is the easy half. The expensive half is deciding which of your own SKUs are in scope, and here the useful metric is margin at risk, not revenue.
Margin at risk is roughly: units sold, times your margin per unit, times how price-sensitive that line is. A high-revenue product where you are the only stockist in the country has almost no margin at risk — nobody is going to undercut you. A mid-revenue product sold by four rivals within two percent of each other has enormous margin at risk, because a single competitor move changes your conversion rate that week.
A worked cut
A mid-size electronics retailer, roughly 9,000 SKUs. This is the shape of the exercise, not a template to copy.
| Segment | SKUs | In scope? | Why |
|---|---|---|---|
| Top sellers, 3+ rivals stock them | 180 | Daily | This is the margin at risk. Everything else is a rounding error against it. |
| Top sellers, sole stockist | 60 | Monthly | Nobody to be undercut by. Checked only to catch a new entrant. |
| Long tail, any rival | 6,200 | No | Individually immaterial; collectively would be 90% of the bill. |
| New lines, first 90 days | ~120 rotating | Daily | Launch pricing is where mistakes are cheapest to catch and most expensive to leave. |
| Own-brand / exclusive | 2,400 | No | No comparable listing exists, so there is nothing to match against. |
Do the arithmetic before you commit
Your daily page count is competitors times in-scope SKUs that each competitor actually carries. Not competitors times your whole catalogue — a rival who stocks 40% of your in-scope range only contributes 40% of those pages.
For the example above: 180 daily SKUs, five competitors, average carriage of about 60%, gives roughly 540 pages a day, or about 16,000 a month. That is a number you can take to whoever signs off the budget, and it is small enough that you can afford to check it twice a day later if the business case justifies it.
Now run the same sum for the "track everything" version: 9,000 SKUs across five competitors is in the region of 800,000 pages a month. Same project, fifty times the bill, and a report nobody can read.
Worked example: ranking four lines by margin at risk
Margin at risk is monthly margin times the share of it a cheaper competitor could plausibly take. Revenue puts these four lines in the order A, C, D, B. Margin at risk puts them in the order C, A, D, B — promoting the second-biggest line by revenue to the top, and demoting the fattest-margin product in the business to nothing at all, because nobody else sells it.
| Line | Monthly revenue | Monthly margin | Share a cheaper rival could take | Margin at risk |
|---|---|---|---|---|
| A. Branded 55-inch TV, every rival stocks it | €180,000 | €7,200 | 90% | €6,480 |
| B. Own-brand HDMI cable | €36,000 | €24,000 | 0% | €0 |
| C. Branded coffee machine, three rivals stock it | €110,000 | €20,000 | 70% | €14,000 |
| D. Sole-stockist power tool | €60,000 | €18,000 | 10% | €1,800 |
What usually goes wrong
The selection stage has no error messages, so these are only ever discovered through the bill or through a report nobody trusts.
- Scoring competitors on how often the sales team mentions them. Irritation is not overlap. Twenty minutes of clicking through your top 100 is the only input here that survives scrutiny.
- Letting the list grow by one competitor per meeting. Every name added is a permanent multiplier on every future run, and nobody ever proposes removing one.
- Tracking own-brand lines because they are top sellers. They cannot be matched to anything, so they generate pages, cost money and return nothing — line B above is the whole argument.
- Treating one marketplace as one competitor. A single listing page can carry fifteen sellers, and the price that matters is whoever currently holds the buy box, not the marketplace's own.
- Pricing the project on the SKU count rather than the SKU-times-competitor count, and discovering the real figure after the budget is signed off.
Write these down before you move on
Lesson three assumes you have them.
- The named competitor list, split into daily and less-than-daily
- The in-scope SKU list, with the rule you used to build it written next to it
- Your estimated monthly page count, and therefore your estimated monthly cost
- The review date — scope rots, and a list nobody has revisited in a year is tracking last year's market
You have a list of competitors and a list of products. Next you need the thing that connects them: the actual URLs of their product pages.
Lesson 3: finding competitor product URLs