How to Calculate a Competitive Pricing Index

Most of what has been written about a price index concerns inflation. A consumer price index follows one basket of goods through time to say how much a currency has lost, which is a market compared with its own past. The number on this page does something unrelated: it compares your price with other companies’ prices, at one moment, for one thing a buyer is choosing between.

The method was built for identical products

The competitive version comes from grocery and ecommerce, and it works beautifully there for a reason worth spelling out. Two shops sell the same barcode. The product is identical, the unit is identical, and the only variable left is the price, so dividing one by the other is a measurement and the result is not open to argument.

Business software has no barcode. Two products in the same category have different features, different limits, different things included, and frequently a different unit of sale altogether. There is no identical item to compare, so the comparison has to be built before it can be made. That single difference is what separates a defensible index from a decorative one, and it is why almost everything below concerns the word matched rather than the division.

What the number is actually for

It answers a question a pricing teardown answers too slowly: where do we sit, in one figure, against the companies a buyer is genuinely choosing between. That makes it a reporting and tracking instrument rather than an analytical one. It goes on a quarterly slide, it moves when the market moves, and it gives a pricing conversation a starting point that is not somebody’s impression of whether the sales team complains about price.

Your price on top, and never the other way round

Both conventions exist in the wild, and a series where somebody quietly swapped them is unreadable, because 117 and 85 describe the same position from opposite ends. Put your own price in the numerator so that above 100 means you cost more, write that sentence into the report header, and the ambiguity never arises again.

The median, and at least three competitors

A median needs three values before it behaves, and it is the right summary here because shortlists routinely contain one vendor priced far away from the others. A mean lets that vendor move your index several points on their own, which reports on them rather than on you.

The set is the other decision nobody documents. Include the companies a buyer actually shortlists against you, which is usually three to five names and is knowable from your own deal records rather than from an analyst’s category. Adding vendors nobody compares you with pulls the figure towards a market you do not sell in. Name the set in the header and keep it stable, because changing it changes the number.

Everything difficult about a competitive pricing index happens before the division. A configuration is a written specification of exactly what is being priced, applied identically to every company in the set, and an index without one is arithmetic performed on unlike things.

1. 1The unit of sale. Seats, usage, tracked records, a flat platform fee. Where yours and theirs differ, the index stops being a single number, which is the subject of the next section.
2. 2The volume. A seat count or usage level, stated. This is the input that changes the answer most and the one most often left implicit.
3. 3The term and billing cadence. Annual billing is routinely cheaper than monthly by the equivalent of a month or two, so comparing your annual rate against their monthly one manufactures an advantage nobody has.
4. 4The capability set. The functions a buyer in this configuration needs, priced wherever each company gates them. If a capability sits in your middle tier and their top one, the honest comparison prices their top tier.
5. 5What is charged separately. Implementation, mandatory support, minimum seat counts, overage rates, required add-ons. These sit outside the headline price and they are where a comparison most often flatters whoever the analysis belongs to.
6. 6The date and the currency. One date for every price in the set, one currency, one region. Prices collected across three weeks and two regions produce a number nobody can reproduce.

The first item on that list is the one to resolve first and the one that takes real work. The pricing teardown template works through identifying what a competitor’s bill actually scales with, along with the gates and the fees that never appear on a pricing page. Do that once per competitor and the index becomes a quarterly arithmetic job rather than a research project.

Where the prices themselves come from is a separate question with its own answers. Finding competitor pricing sets out the published sources and grades their reliability, which matters here because an index is only as reproducible as its least documented input.

This is the part retail never had to think about, because a tin of tomatoes does not come in tiers. Once either company prices in bands, or the two of you sell in different units, the ratio between your prices changes with every customer. The index is then a function of who is buying, and quoting one value for it hides more than it reports.

Take a simple and entirely ordinary pair. You charge $49 per seat per month. One competitor charges a flat fee by tier: $900 up to 20 seats, $2,400 up to 75, $6,000 up to 200.

Not one price changed between those rows. The same two price lists produce an index of 54 and an index of 163, and the figure does not even move in one direction: it climbs, falls back, climbs again. Any single number claimed for this pair is a statement about which customer somebody happened to pick.

The index crosses parity three times

Work out where the two prices meet and you get three answers rather than one. At 18 seats the index is 98 and at 19 it is 103, so parity falls between them. Inside their second tier it crosses again at 49 seats, and inside their third at 123. Below about nineteen seats you are reliably the cheaper option, and from a hundred and twenty-three up to the top of their published range you are reliably the dearer one. In between, the answer alternates band by band, depending on which side of one of their tier boundaries a customer happens to fall.

Report bands, not a figure

What survives contact with a pricing meeting is three numbers and a sentence: the index at a small, a typical and a large configuration, plus where parity falls. That takes one extra row in a spreadsheet and it answers the question the single figure only pretends to. It also tells marketing something specific, which is the customer size at which a claim about being the affordable option stops being true.

Everything above uses published prices, and published prices are not what most business software sells for. That does not make the list index worthless. It makes it one of two indices, and the two answer different questions for different people.

What each one decides

The list index decides whether you reach the shortlist at all. Buyers compare published pricing pages long before they speak to anybody, and a figure well above 100 removes you from consideration by people you will never meet and cannot persuade. It is a marketing number, and its audience is a stranger with a browser open.

The street index decides whether winning is worth it. Built from what buyers actually signed, it reports on the deals you competed for rather than on the ones you were considered for, and its audience is whoever owns margin. A company can be comfortably above 100 at list and below it at street, which means it looks expensive to everybody who has not talked to it and is quietly the cheap option to everyone who has.

Where the second index comes from

Your own transacted prices are exact and sitting in a system. Theirs are the hard half, and they come from your own closed-lost records, from published contract awards, and from asking the buyers who evaluated both of you what the other quote actually said. A handful of deals is not a price list, so this second index firms up slowly and is worth rebuilding only once enough have closed to move it.

One caution about attributing the gap. If your street index sits well below your list index, the cause may be that you discount harder rather than that they hold firm, and those call for opposite responses. Discount rate measures your own side of that, against your own list rather than against anybody else, and reading the two together is what separates a pricing position from a negotiating habit.

Everything above is a set of decisions, and a decision nobody wrote down gets remade differently next quarter. The artefact that prevents that is small: one sheet, rebuilt on a fixed cadence, carrying its own assumptions in the same file as its answer.

One sheet, one row per configuration

Columns for the configuration, the date, each company’s price, which tier that price came from, what it includes, the median, your index, and a note for anything that changed since the last run. Three rows, one per configuration band. Every one of those columns exists because somebody challenged the figure once and the answer was not written down.

The prices come out by hand from published pages, marketplace listings and, where they exist, contract award records. There is no export for a competitor’s price list, which is why the dated screenshot filed beside the sheet is part of the method rather than a nicety: a price you cannot evidence is a price somebody will dispute in six months.

What a connected system changes here

The arithmetic is not the cost of this metric. Rebuilding the inputs every quarter is, and that is the part a competitive intelligence platform removes, by keeping a current record of each competitor’s plans, prices and inclusions so the quarterly run is a division rather than a fortnight of re-reading pricing pages. The configuration still has to be decided by a person, once, and then held.

There is a failure specific to this metric that has nothing to do with anyone changing a price. A competitor moves one capability from their middle plan to their top plan, and not a single figure they publish changes. What moves is the denominator of your index: the configuration you matched now buys less from them than it did last quarter, so the two things being divided have stopped describing the same purchase, and the metric carries on producing a number that looks entirely reasonable.

A price cut announces itself. Anybody watching the page sees it, sales hears about it inside a week, and the index gets corrected. A repackaging is silent by construction, and it is the more common event of the two. That asymmetry is why this metric goes wrong slowly and why the version most companies hold is more confident than it has earned.

Catching it means watching what each tier contains rather than what it costs. Flares follows competitor plans, prices and what is included at each level, and puts a date on every change, so a capability crossing a tier boundary surfaces as an event instead of as an index nobody can explain two quarters later.

What no tool decides for you is the configuration. Which capabilities a buyer in your segment actually needs is a call only somebody close to your deals can make, and it should be revisited deliberately rather than drifting. What monitoring supplies is the evidence that the configuration you chose still means what it meant when you chose it.