How to choose
pricing software

Three vendors, four discovery calls each, two RFI rounds, six demos, and a 47-tab feature comparison spreadsheet that nobody on the buying committee has actually read end to end. By the time you reach the final round, the decision is being made on which CSM smiled most. This is not because pricing managers are bad at evaluation — it's because the way most evaluations are structured selects for the wrong signals.

The way out isn't more features in the comparison. It's a smaller number of higher-resolution criteria, with the right weight, and the right vendor question for each one. Below is the framework we use with mid-market non-food retailers in active evaluations. Twelve criteria, prioritised. Forty-plus questions designed to surface real differences. One realistic timeline.

Part one · The 12 criteria

What actually
separates platforms

Each criterion below has a priority signal: Must-have (deal-breaker if missing), Important (significant impact, but recoverable), or Nice-to-have (relevant but rarely decisive). Use the weight column when building your scorecard — and always have one column for "evidence the vendor showed" alongside the score.

Two notes on using this. First, weights are starting points, not gospel — adjust based on what's broken in your business today. If you have a fragile ERP integration, push integration depth to 15%. If your team is small and senior, lower workflow editor weight. Second, never let a single vendor score below 3 on a Must-have criterion. Below 3 means walking away, regardless of total score.

The most useful question across all twelve isn't on the list: "Show me, on your platform, with our data, in a screen-sharing call." Vendors who can do this in week one of evaluation are different from vendors who need three weeks of data prep. The gap shows up in everything downstream.

Part two · The timeline

What a six-month
evaluation looks like

Pricing software evaluations almost always take longer than expected. Plan for six months end to end. Compressing it works against you — vendors will rush demos, your team won't get to live test, and you'll sign on incomplete information. The phases below assume serious evaluation, not informational shopping.

The pattern that breaks evaluations: skipping proof-of-concept. Vendors will tell you "the demo is the proof" or "we have so many references you don't need a POC." Both are usually true, and both miss the point. The POC is not for the vendor to prove they work — it's for your team to learn the platform under realistic conditions before committing. The 4–6 weeks you invest pay back many times over in the implementation that follows.

Part three · Walk-away signals

When to leave
the evaluation

Six patterns that should end an evaluation regardless of how strong the rest of the pitch is.

• Implementation timeline that sounds impossible. "Live in two weeks" with a serious mid-market non-food retailer means thin product, thin implementation, or both.
• AI claims the technical team can't explain. If the data scientist on the call can't describe the model running the pricing decisions, it's marketing, not engineering.
• No customer references at your scale and complexity. Enterprise references don't translate down; small-business references don't translate up. You need same-segment proof.
• Pricing structure that punishes growth. Per-SKU pricing that triples when your assortment doubles is the cost equivalent of a referendum on whether you should expand.
• Evasive on data exit. If the answer to "how do we leave" is anything other than a clear price and timeline, vendor lock-in is real.
• Discovery call that asks for almost no detail about your business. A serious vendor wants to qualify out fits as much as in. The ones that don't are selling a generic product to anyone who'll buy.