The default selection rule for testimonials in most marketing teams is "is it enthusiastic?" The five-star praise gets published, the lukewarm-but-specific quote gets archived, and the long thoughtful customer reflection that is six paragraphs long gets discarded as "too long." This rule produces the wall of "Game-changer for our team" quotes that lines so many SaaS landing pages and that converts almost nothing.
The fix is not to look harder for better testimonials. It is to apply a deliberate scoring rubric to whatever submissions arrive, with explicit thresholds for publish / follow-up / archive. The rubric below is the one we use internally — five criteria, each scored 0-3, with cut-offs at total scores of 12 (publish), 8-11 (request a follow-up), and below 8 (archive without further work).
Why a rubric and not a vibes call
Three reasons a written rubric beats a vibes-based selection process even when the marketer making the call is experienced:
Repeatability. A rubric produces consistent decisions across reviewers and across time. A vibes call from the same marketer on Monday and Friday can land in different places, and a rubric collapses that variance.
Defensibility. When the founder asks why testimonial X did not make the landing page, a rubric gives you a specific answer ("scored 9, in the follow-up range") instead of a defensive "it didn't feel right."
Volume scaling. As the pipeline of submissions grows, a vibes process becomes a bottleneck on whoever has the taste. A rubric can be delegated, distributed, or partially automated.
The five criteria
Each criterion scored 0-3. Total score 0-15. Cut-offs as stated above.
1. Specificity (0-3)
What concrete fact, number, or scenario does the testimonial reference?
- 0 — Pure adjective stack ("Amazing product, love it, can't recommend enough"). Zero concrete content.
- 1 — One mild specific reference ("Saved me time on customer support").
- 2 — A clear scenario or workflow described ("Used to spend 20 minutes per testimonial publishing it manually, now it's two clicks").
- 3 — Numbers, time-frames, before/after comparison ("Went from a 2.1% to a 4.4% landing-page conversion in eight weeks after rolling out testimonial mapping to our top three objections").
Specificity is the single highest-impact criterion. A 3 on specificity will outperform a 3 on every other criterion combined for landing page conversion. Most submissions land at 0-1; chasing the 2-3 cohort is most of the work.
2. Attribution strength (0-3)
How strongly is the testimonial attributable to a specific identifiable person?
- 0 — Anonymous, no name, no role, no company.
- 1 — First name and role only ("Sarah, Marketing Manager").
- 2 — Full name and company name ("Sarah Chen at Aperture Industries"), no photo.
- 3 — Full name, company, role, photo, and (ideally) a link to a public profile.
Attribution strength governs whether the reader's brain processes the testimonial as evidence or as marketing copy. Anonymous testimonials read as marketing copy regardless of how good the content is. The first jump (from 0 to 1) recovers most of the credibility loss; the final jump (to 3) provides resilience against skeptical readers in B2B contexts.
3. Objection coverage (0-3)
Does the testimonial address one of your top buyer objections?
- 0 — Generic praise that does not map to any objection on your list.
- 1 — Vaguely touches on a category that includes one of your objections.
- 2 — Directly addresses a single objection on your top-five list.
- 3 — Addresses one of your top-three objections and resolves it concretely.
This criterion is the backbone of objection-mapped placement on landing pages — see the objection-handling testimonials guide for the full placement framework. A testimonial scoring 3 here can lift conversion noticeably even when its specificity score is mediocre, because the placement-by-objection effect dominates.
4. Recency (0-3)
How recent is the testimonial?
- 0 — More than 24 months old or undated.
- 1 — 12-24 months old.
- 2 — 6-12 months old.
- 3 — Less than 6 months old.
Recency matters because product changes, market context, and feature parity all drift. A two-year-old testimonial about a feature that has since been re-architected is actively misleading. Recency also affects perceived credibility — readers (rightly) discount old testimonials. The testimonial freshness operations are covered in our testimonial rotation and freshness guide.
5. Format fit (0-3)
How well does the format match the placement you have available?
- 0 — Format is wrong for any current slot (10-paragraph essay when you only have inline quote slots).
- 1 — Could fit with substantial editing.
- 2 — Fits a current slot with minor edits.
- 3 — Fits a current slot as-is.
Format fit is the criterion where you can do the most repair-by-editing without integrity loss. A high-content, low-format-fit submission can often be converted to a 3 by trimming to a key sentence — but trimming should be confirmed with the source rather than published silently after the fact.
The scoring workflow
The mechanical workflow we run is:
- Submission lands in the testimonial intake (form, email, support ticket annotation).
- First-pass scoring within 48 hours by whoever owns the rubric. Five numbers, total recorded.
- Routing by total:
- 12-15: publish-eligible. Format check, attribution permission confirmed, scheduled into the next refresh slot.
- 8-11: follow-up requested. Specific clarifying question that targets the lowest sub-score (e.g., "Could you share the rough percentage improvement?" if specificity is 1).
- 0-7: archive. Thank-you note sent, submission stored in case the customer's situation evolves.
- Re-score after follow-up. Submissions that came back from the follow-up loop go through scoring again before publishing.
The two failure modes of any rubric
Rubrics fail in two specific ways and you should watch for both:
Score inflation. Over time, reviewers drift toward giving higher scores because cuts feel mean. A rubric with average score 11 across all submissions has stopped distinguishing anything. Periodically re-calibrate by reviewing the bottom-quartile of last quarter's published testimonials and asking whether they would still score above the 12 threshold today.
Criterion neglect. Reviewers focus on the criterion they personally weight highest (often specificity) and rubber-stamp the others. Reviewing the standard deviation of scores across criteria is a quick health check — if four criteria have nearly the same variance and one has near-zero variance, the low-variance criterion is being neglected.
Where the rubric does not apply
Two cases where you should not run the rubric:
- Outsized story testimonials. A customer whose story is genuinely remarkable (10x business outcome, dramatic before/after) is doing different work than a routine testimonial. They get a case study, not a quote slot, and the rubric is the wrong instrument.
- Press quotes from publications. "Recommended by TechCrunch" is doing brand-association work, not customer-evidence work, and is scored on a separate axis. Do not stuff press into the customer-testimonial pool.
For everything else — the bulk inflow of customer quotes that arrives via email, support tickets, NPS follow-ups, and submission forms — the rubric is the operational instrument. Run it weekly, audit it quarterly, and use the inflow data to inform what objections you should be requesting follow-ups about. The intake-to-publish pipeline is detailed further in our testimonial collection automation workflow guide.
Closing note
The shift from vibes-based to rubric-based testimonial selection is a small operational change with a disproportionately large output effect. Most teams discover, on first pass, that 30-50% of what they have been publishing scores below their own publish threshold once they have written one down. The cleanup of that backlog plus a tighter forward filter is typically the highest-ROI hour of work the marketing team will do that quarter.