Competitor price analysis: what to do once you have the data

Competitor price analysis is the work that happens after collection: turning a file of competitor prices into a handful of measures a pricing team can act on. Seven metrics carry almost all the value. Price index, position mix, win rate by category, gap distribution, volatility, promotional cadence and margin at risk. Each answers a specific question and points at a specific decision.

Competitor price analysis starts where collection ends

The common failure is not missing data. It is a full price file and no answer. An analyst exports 3,000 rows, builds a lookup, makes three pivot tables, and the meeting still turns into anecdotes about two SKUs somebody checked by hand. More data does not fix that. A fixed set of metrics, computed identically every month, does: once the definition stops moving, movement in the number starts to mean something.

Two rules make everything below trustworthy. Exclude out-of-stock listings from every market calculation, because a price nobody can buy is not a price you compete with. And compare only confirmed matches. An index built by comparing a $199 blender against a $59 accessory is worse than no index, because it is confidently wrong and somebody will act on it.

The seven metrics of a competitor price analysis

Price index, with the arithmetic shown

price index = (our price / mean of in-stock competitor prices) x 100

  100   = at market
  < 100 = cheaper than the market average
  > 100 = more expensive than the market average

Take a single SKU. Our price is $199.00. Four competitors are in stock at $189.00, $205.00, $212.00 and $198.00. Those sum to $804.00, so the mean is $201.00. Divide 199.00 by 201.00 to get 0.9900, times 100, and the price index is 99.0. We are one percent below the market average. If a fifth competitor is listed at $149.00 but out of stock, including it drags the mean to $190.60 and the index to 104.4, which reads as expensive against an offer nobody can buy.

Rolling a SKU index up to a category

A category index is not the average of its SKU indexes, or it lies to you. Three products, as an example. A sells 500 units a month at index 96, B sells 60 at index 118, C sells 40 at index 121. The unweighted mean is (96 + 118 + 121) / 3 = 111.7, which reads as badly overpriced. Weight by units and you get (500 x 96 + 60 x 118 + 40 x 121) / 600, which is 59,920 / 600 = 99.9. You are at market where the volume is and expensive on the tail. The first number sends you into a broad price cut you do not need. The second sends you at two SKUs.

Weight by units when you are protecting share, by revenue when you are protecting margin dollars, and report both when they disagree by more than a point, because that disagreement is itself the finding. The price index and price positioning guide goes deeper on weighting.

Price position mix: whether you are actually the cheapest

Price index is an average, and averages hide their own shape. Position mix does not. Bucket every SKU against the lowest in-stock competitor: lowest if you beat it, matched if you are within a tolerance such as fifty cents, above if you are not. Then count the buckets.

Example. A catalog of 240 tracked SKUs comes back lowest on 61, matched on 44, above on 135. That is 25.4 percent lowest, 18.3 percent matched, 56.3 percent above. A team describing itself as the value option is the value option on a quarter of its catalog. That one number changes what marketing is allowed to claim and shows how much of the range carries it.

• Run the mix on the revenue top 50 as well as the whole catalog. Shopper perception is set by the products people actually compare, and those are rarely spread evenly through an assortment.
• Watch the matched bucket. A large matched bucket usually means several competitors are running repricers against each other, and margin is leaking out of the category for nobody's benefit.
• Split the above bucket by how far above. Two percent above the market low is a defensible position. Eighteen percent above, with no service or bundle attached, is an invisible product.
• Track the mix month over month, not day to day. Daily movement is promotional noise. Test a scenario first with the price position analyzer.

Win rate by category is the same calculation cut differently: the share of tracked SKUs in each category where you are at or below the lowest in-stock competitor. It settles the argument about whether pricing is a company-wide problem or a category-specific one. It is usually category-specific, and the worst categories are usually the ones nobody has repriced since launch.

Price gap distribution beats the average gap

The average price gap is the most comforting useless number in pricing. Five SKUs at gaps of -12%, -1%, 0%, +2% and +11% average to exactly 0.0 percent, so the report says you are at market. In reality one product gives away twelve points of price and another is eleven points too expensive. The errors cancel on the summary and both keep costing money.

Look at the distribution instead. Sort every SKU by its percent gap to the market average, then read the median and the two tails. The median tells you where the catalog really sits, often nowhere near the mean. The bottom decile is where margin is left on the table, the top decile is where you are invisible, and on a 3,000 SKU catalog each decile is 300 products: a real week of work with a real result.

1. 1Bottom decile first. You are cheapest here by a wide margin. Raise toward the market low in small steps and watch units, because this is the fastest margin available in the whole analysis.
2. 2Top decile second. Decide product by product: cut to the market, or accept that the product is not price-competitive because it is bundled, exclusive or serviced differently.
3. 3Leave the middle alone for now. SKUs within a couple of points of the market average rarely repay attention until the tails are dealt with.
4. 4Re-read the tails after every pass. They refill, which is why this is a monthly routine rather than a project.

Volatility and promotional cadence: knowing when to look

Volatility is price changes per SKU per 30 days, computed per category. It is not interesting on its own. It is the metric that sets every other cadence you have. Suppose Audio changes 6.2 times per SKU per month while Home and kitchen changes 1.1. That is roughly one change every five days against one every four weeks. The same refresh rate cannot be right for both, and neither can the same review cycle.

Promotional cadence is the shape inside the volatility: the same discount depth at the same interval. A competitor who drops fifteen percent for four days at the start of every month is running a calendar, not reacting to you, so plan around it rather than match it within the hour. Sixty to ninety days of daily history is usually enough to see the pattern; under thirty days you are looking at coincidence.

Margin at risk: the number that gets the decision made

Everything above tells you where you stand. Margin at risk tells you what a move would cost, which is what settles the argument. One SKU, worked through. Cost is $120.00 and our price is $199.00, so unit margin is $79.00 and gross margin is 39.7 percent. The lowest in-stock competitor is at $179.00. Match it and unit margin falls to $59.00, a drop of $20.00, or 25.3 percent of the margin dollars on every unit. At 400 units a month, matching costs $8,000 a month in gross margin.

Now the decision is concrete: is matching worth $8,000 a month, and how many extra units would break even? At $59.00 of margin per unit you need roughly 136 extra units a month to stand still, a 34 percent volume lift. Write that down before the meeting and most match-the-market debates end in ninety seconds. Sum the same calculation across every SKU where you sit above the market low and you have the portfolio number: margin dollars exposed if you matched everything.

Break-even volume is where analysis hands off to demand. The margin arithmetic is mechanical and runs in the margin calculator. Whether the volume actually arrives is an elasticity question, covered in the price elasticity guide.

A monthly competitor price analysis agenda you can copy

Sixty minutes, the same agenda every month, one owner per decision. The timeboxes matter more than they look: without them the meeting spends forty minutes on the first interesting SKU and never reaches the portfolio view.

1. 1

   Data quality (5 minutes)

   Coverage and freshness by source since the last review. Any source below threshold gets fixed before its numbers are discussed. If a source was degraded during the month, say so now rather than mid-argument.

2. 2

   Index movement (10 minutes)

   Overall index and index by category versus last month, weighted by units or revenue. Explain any move over about 1.5 points. Most moves are a competitor promotion rather than anything you did, and naming which is which is the whole exercise.

3. 3

   Position mix on the revenue top 50 (10 minutes)

   Lowest, matched and above, this month against last. This is the slide that tells you whether the price perception you pay for in marketing is supported by the price you actually charge.

4. 4

   The tails (15 minutes)

   Bottom and top decile of price gap, by name. Each SKU leaves the room with a decision, an owner and a review date, or is explicitly parked. Nothing is left as needs more analysis, which is where SKUs go to be forgotten.

5. 5

   Competitor behavior (10 minutes)

   New sellers detected, cadence changes, categories where volatility jumped, and any competitor that appears to have switched on automated repricing. That last one changes how fast you react all year.

6. 6

   Margin at risk and the match list (5 minutes)

   Total margin exposed if you matched everything you are above on, then the short list you will actually match, with break-even volume next to each one.

7. 7

   Rules, not one-offs (5 minutes)

   Turn every repeatable decision into a repricing rule with guardrails so it needs no human next month. Anything that cannot be expressed as a rule stays a manual item with a named owner.

Turning the analysis into prices

Analysis that never changes a price is a hobby. The handoff is a rule with guardrails: a scope, a strategy, hard floors for margin, cost and MAP, and a cap on how far a price moves in a day. Changes land in a review queue, export as CSV, or push straight to a connected store, and every change is kept with its before and after. That is repricing, and repricing software covers how the rules are built.

Two honest limits. This analysis tells you where you stand and what a move would cost; it does not hand you the profit-maximizing price, because that needs demand response you have to measure. And it cannot tell you why a competitor is cheaper. A lower cost base, a loss leader, a clearance lot and a pricing mistake look identical in a price file, which is why the analytics module is meant to be read next to somebody who knows the category.