Benchmarks and percentiles

A benchmark is a distribution, not a number. Take every comparable deal, deals of similar size, product mix, segment, and vintage, and line up their net prices from best to worst. The market low is the strong end of that line, the median is the middle, and the market high is the weak end. Your deal lands somewhere on the line, and where it lands is your percentile.

Comparability is what separates a benchmark from a rumor. A price a 50,000 seat bank pays tells a 500 seat company very little, because they sit in different discount bands. Good benchmarks control for deal size, product edition, region, and when the deal closed, since street pricing moves over time. When someone quotes you a market price, the first question is always the same: comparable to what.

One direction note that trips people up: percentiles are stated so that higher is better for you, the buyer. The 75th percentile on discount means a stronger discount than about three quarters of comparable deals. On a cost-per-unit metric the arithmetic flips, lower cost is better, but a well built benchmark normalizes this so that a high percentile always reads as a strong position.

The median is the pivot. Half of comparable deals landed better than the median and half landed worse, so the median is what an informed buyer with ordinary leverage achieves. If you are below it, you are subsidizing the buyers above it, and moving to the median is not an ambitious ask, it is a correction. Vendors know this, which is why a well evidenced move to median is the easiest negotiation win available.

Above the median, the conversation changes. Between the 50th and 75th percentile you are doing well, and further gains come from trading things the vendor values: term length, timing, references, growth commitments. Past the 75th percentile you are in the territory of buyers with unusual leverage, and pushing for the absolute market low without that leverage burns goodwill you may want at the next renewal. Strong analysts aim for a defensible position, not a trophy.

Also read the spread. When the low and the high sit close together, pricing is commoditized and there is little room to negotiate on price, so shift your effort to terms. When the spread is wide, discounting is discretionary and preparation is decisive, because the difference between an informed buyer and an uninformed one can be 30 points or more on the same product at the same volume.

A benchmark becomes leverage when you convert it to currency. The sentence that moves a negotiation is not we feel the price is high. It is comparable deals at our volume land at a median of X, we are currently at Y, and the gap is costing us Z per year. Naming a credible number changes the vendor's math, because now they are defending a measurable gap rather than debating a feeling.

Use the percentile to set both ends of your range. Target the band between the median and the 75th percentile as the outcome, and treat your current position as the floor the vendor must improve on. Anchoring your first ask above the target is standard, but keep every number inside the range the data supports. An ask the data cannot support invites the vendor to dismiss the data, and the data is your leverage.