How to Measure a SaaS Launch Experiment
A practical protocol for comparing launch channels using time, costs, qualified signups, and verifiable attribution.
By SaaS Hunt Editorial · Reviewed 2026-10-07
A protocol, not a results report
No completed launch experiment is claimed here. Use this protocol to collect evidence before publishing a case study.
1. Define a useful outcome
Pick one product and define a qualified signup or activation before launch. For a directory, a reviewed product submission is more useful than an account creation or an outbound click. Record what your analytics can and cannot measure.
2. Select three channels
Choose audiences that plausibly need your product. Keep the landing page and product offer consistent. Record timing, existing audience size, and any promotion that might affect comparison. Check the official rules for each channel.
3. Keep a launch log
Record the published URL, campaign link, publication time, hours spent, cash cost, and currency. Use a separate utm_source per channel and one shared campaign name. Check that the link preserves attribution.
4. Review at 7 and 30 days
Record sessions, qualified signups, activations, and costs for each fixed window. Report cost per qualified signup only when signups are nonzero. Show raw counts alongside rates, and leave missing data blank rather than calling it zero.
5. State what the data cannot prove
Different launch dates, returning visitors, blocked analytics, and small samples limit comparison. GA4 session attribution is not a randomized experiment. A launch platform’s outbound links are different from its users voluntarily linking to you.
6. Publish a reproducible account
Explain the product, dates, channel choices, definitions, results, and mistakes. Use aggregate screenshots with private data removed. Separate direct measurements from estimates, and acknowledge unsuccessful channels.