How to build your own MMP shortlist

Start with your constraints

A shortlist is the result of cutting, so you need something to cut with. Before you open a vendor page, write three lines.

• Channels. Every ad platform you buy from now, and any you plan to add within a year.
• Budget. The most you will pay per month for measurement, and whether you can sign for a year.
• People. Who installs the SDK, who builds links, who reads reports. On a small team this may be two people.

These three lines decide more than any feature grid. A team with one developer and four channels has a different shortlist from a team with a data engineer and twenty.

The five steps

1. Collect the long list. Take names from a published list such as the six on our home page, from founders you know, and from the partner pages of the ad platforms you use. Do not judge yet.
2. Write three or four must-haves. Each one must be a yes-or-no test that follows from your constraints. "Connects to all my current channels" is a test. "Good reporting" is a wish.
3. Cut on public facts. Check each vendor's own plan and pricing pages against your must-haves. If a page does not answer the test, mark it unknown and send one email. Reading a plan page line by line helps here.
4. Rank what is left on nice-to-haves. Now the softer points count: report layout, documentation quality, how fast support replied to your email.
5. Trial the top two or three with real traffic. A trial with no ads running proves little. Know what to verify in a trial before day one.

Must-haves that work as tests

The table shows how a vague need becomes a test you can settle from a vendor's own pages.

Apply every test to the plan you would actually buy. A feature on a product page may sit in a higher tier, a pattern covered in capabilities often outside an entry plan.

Mistakes that make a shortlist useless

Scoring before cutting. A weighted spreadsheet with twenty criteria lets a vendor that fails a must-have win on points. Cut first, score second.

Listing features you will not use this year. Planning for an imagined future team makes every enterprise plan look necessary. Choose for the next twelve months and check that moving up or moving out is possible later.

Treating a demo as a trial. A demo shows sample data in a tidy account. Your own SDK install, links and campaigns are the only real test.

Forgetting what the tool leaves out. If your real question is about product behaviour or messaging, read what an MMP does not do before you compare any vendors at all.

What to do next

When you are down to two or three names, send each one the same ten questions so the answers can be compared side by side.