# Exit Interview Vendor Scorecard Score any exit interview or exit survey vendor against the six things that actually determine whether you get usable churn insight. From saasfeedback.ai — https://saasfeedback.ai/blog/exit-interview-software-saas --- ## How to use it Score each criterion 0–5. Multiply by the weight. Anything under 60/100 will produce data you do not act on. Run the same sheet across three vendors before you talk to any of them about price. | # | Criterion | Weight | Score (0–5) | Weighted | |---|---|---|---|---| | 1 | Trigger coverage | 20 | | | | 2 | Reach mechanism | 20 | | | | 3 | Response depth | 25 | | | | 4 | Analysis output | 15 | | | | 5 | Routing & loop closure | 10 | | | | 6 | Time to first insight | 10 | | | | | **Total** | **100** | | | --- ## 1. Trigger coverage (weight 20) Which departure events does it actually fire on? - **0** — Cancellation only, and only if the user clicks through your cancel flow. - **3** — Cancellation plus trial expiry. - **5** — Cancellation, trial expiry, downgrade, dunning failure, and silent inactivity, each with its own configurable delay. *Why it matters:* voluntary cancellation is the smallest and least representative slice of your churn. The users who ghost never touch a cancel flow, and they are usually the ones with the most damning answer. **Ask the vendor:** "Show me the event list. Which of these can I trigger from a Stripe webhook without engineering work?" --- ## 2. Reach mechanism (weight 20) How does it get in front of the departing user? - **0** — In-app modal only. Invisible to anyone who already stopped logging in. - **2** — Email only. - **4** — Email plus SMS, with sequencing. - **5** — Multi-channel including a real human phone call. *Why it matters:* reach caps everything downstream. The best question design in the world scores zero if nobody is listening. **Ask the vendor:** "What is the median response rate across your customer base, by channel? Not best case — median." --- ## 3. Response depth (weight 25) What comes back? - **0** — Multiple choice only. You get counts of your own hypotheses. - **2** — Multiple choice plus an optional free-text box. - **4** — Structured open questions with automatic follow-up prompts. - **5** — Conversation transcripts with real follow-ups, where the second and third question adapt to the first answer. *Why it matters:* this is the criterion vendors are weakest on and it carries the most weight for a reason. "Too expensive" chosen from a dropdown is not a finding. It is the absence of one. **Ask the vendor:** "Show me three real anonymised responses from the last month. Not a case study — raw output." --- ## 4. Analysis output (weight 15) What does it do with the answers? - **0** — CSV export. You do the work. - **3** — Automatic theme tagging with a fixed taxonomy. - **5** — Theme tagging you can retrain, weighted by the revenue behind each theme, with the original verbatims still attached and reachable. *Why it matters:* a theme without its verbatims does not survive a roadmap argument. A theme without revenue attached gets prioritised by whoever speaks loudest. **Ask the vendor:** "Can I see the MRR behind a theme, and can I click from a theme back to the exact sentence that produced it?" --- ## 5. Routing and loop closure (weight 10) Where does the insight land, and what happens next? - **0** — A dashboard someone has to remember to open. - **3** — Scheduled email digest. - **5** — Push into the tools your team already lives in (Slack, your issue tracker, your CRM), with an owner assigned per theme. *Why it matters:* the failure mode of every feedback programme is not collection. It is that the insight arrives somewhere nobody looks. --- ## 6. Time to first insight (weight 10) How long from signing to a finding you could act on? - **0** — Over 8 weeks (implementation project, professional services line item). - **3** — 2–4 weeks. - **5** — Under 1 week. *Why it matters:* long implementations kill feedback programmes before they prove value, and the person who championed the purchase is often gone by the time it lands. --- ## The two questions that end most sales calls early 1. **"Is this software I operate, or a service you operate?"** Both are valid. Confusing them is how teams end up owning a tool nobody has time to run. 2. **"What percentage of my churned users will you get an answer from — and what happens to the ones you don't reach?"** If the answer is "they don't respond", the vendor has sold you a form, not a feedback programme. --- ## Red flags - Response rate quoted without a channel or a denominator. - Case studies where the named metric is "insights collected" rather than a business outcome. - HR exit-interview software repositioned for customer churn with the word "employee" swapped out. The question design is genuinely different; ask to see a customer-churn question bank, not a template gallery. - No way to reach the raw verbatim. - Pricing per response, with no cap. Your worst churn month becomes your worst invoice.