The Software Procurement Playbook

The most common procurement mistake is starting with a tool category. "We need a CRM." "We need an ERP." That framing already locked you into a market and a price band before anyone defined the actual problem.

Start with the job. Write three sentences:

1. What outcome are we trying to produce? ("Stop losing 14% of inbound deals because reps forget to follow up.")
2. Who does the work today, and what do they actually do? ("Five reps using two tabs of HubSpot, a Google Sheet, and Slack DMs to track quotes.")
3. What changes if we do nothing? ("Lost-deal rate keeps growing as we scale headcount.")

If you cannot answer those three sentences without naming a vendor, you are not ready to buy. You are ready to investigate.

You should be able to eliminate 80% of the market in 90 minutes of research. The goal of phase 2 is not to find your vendor; it is to throw out everyone who is not worth a demo.

Ask twelve disqualifying questions before you ever talk to sales:

• Does this product deploy in our region without compliance theater?
• Is the API documentation public and current?
• Do they publish a status page with at least 6 months of incident history?
• Is pricing public, or at least bracketed publicly? (Hidden pricing is a tax on your time.)
• Can we export our data in a documented format on day one?
• Does at least one customer of our size publicly endorse them by name?
• Is there a self-serve trial, or only "book a demo"?
• Do they ship on a published cadence, or is the changelog quiet for months?
• Is the support contract scoped, or is it "best effort"?
• Are the integrations we need actually live, or only on a roadmap?
• Do they require a multi-year commitment for the headline price?
• Can the implementation be done by our team, or do we need their professional services?

Any vendor that fails 3 or more is out. Yes, even the famous one. Especially the famous one.

Most RFPs are 40-page PDFs that ask the same 200 questions every vendor has a canned answer for. They optimize for legal cover, not for picking the right tool. Throw that template out.

A useful RFP has six sections, none longer than a page:

1. The job. Copy the three sentences from Phase 1. No edits.
2. The constraints. Budget range, timeline, team size, must-have integrations, regulatory needs.
3. The decision criteria. Three to five things that will actually decide the contract, ranked. (Not "ease of use"; be specific.)
4. The proof we want to see. Two or three concrete demo scenarios pulled from your real workflow.
5. Pricing & exit terms. Ask for 1, 3, and 5-year total cost. Ask for the data export and termination clause.
6. References. Three customers your size, in your region, that we can call.

Send it to the four vendors who survived Phase 2. Give them 10 business days. If they cannot turn around a 6-section RFP in 2 weeks, their implementation will not be faster.

The vendor demo is theater. Their best presales engineer, on their best canned data, in their best-case path. You will learn almost nothing about whether the product fits your business unless you control the room.

Three rules for vendor demos:

• You drive the agenda. Send the demo scenarios from your RFP 48 hours ahead. If they "did not have time to prepare," that is your answer.
• Your operators run the keyboard for the last 20 minutes. Not the salesperson. Your actual end users, doing their real workflow on a sandbox account.
• Score on the spot. A simple 1-to-5 on each demo scenario, captured in a shared doc before the call ends. Memory degrades fast and salespeople know it.

Watch for these red flags:

• "That's on the roadmap" said three or more times.
• The demo environment is a different version than what you would buy.
• The salesperson cannot answer a technical question without a follow-up email from "the team."
• They keep steering back to features you did not ask about.

Enterprise software pricing is theatre too. The list price exists so the salesperson can offer you a discount. The discount you should expect on annual commit, paid up front:

Annual contract value	Realistic discount
< $25K	10–15%
$25K – $100K	20–30%
$100K – $500K	30–40%
$500K+	40%+, plus implementation credits

What to ask for, in order of how much it costs them:

1. Implementation credits or free onboarding (cheap to give, big to receive).
2. Price lock for 24 months on renewal.
3. A real exit clause: prorated refund + data export window + no penalty if they miss an SLA in year one.
4. A written escalation path with a named technical contact, not "support@."
5. The discount on the annual price.

Sign nothing without your team confirming they can actually do the work in the implementation timeline. The most expensive software contract is the one you never use.