Definition
A screening rubric translates an acquisition mandate into repeatable assessments. It specifies what is being asked, which values can be recorded, what evidence is sufficient, and how the result affects eligibility or prioritisation. It should distinguish mandatory conditions from preferences and provide an explicit treatment of unanswered or disputed questions.
A rubric is more than a weighted scorecard. Weights do not define the meaning of a question, and a high score can conceal missing evidence. If a criterion requires contracted maintenance, a vague reference to ongoing customer relationships should not automatically satisfy it. The accepted values and support requirements should make that boundary clear before research starts, rather than being improvised to fit the company under review.
In origination work
Tailor the questions to the client's mandate, test them on varied examples, and review exclusions as well as apparent matches. Docket's agents use the client's criteria as their questions; Scout records typed values with supporting retained-page quotations. Version changes so the team can identify which rules applied to each assessment. Keep completeness and fit distinguishable: a candidate with many unanswered fields should not appear strongly qualified simply because the few known answers happen to be favourable.
Anatomy of a rubric
A screening rubric turns an acquisition mandate into a fixed set of questions. Each question has a stated meaning, a list of accepted answers, the evidence required for each answer, and a rule for what to do when evidence is missing. For example: "Does the company operate in the state of X? Accepted values: yes, no, unknown. Evidence: a registry record or a page naming a location." The fixed answer list is what makes results comparable.
Decades of research on prediction support this approach. Meta-analyses comparing clinical judgement with mechanical combination of the same information found that mechanical methods were at least as accurate in the large majority of studies, and work on improper linear models showed that simple equal-weight rules often perform surprisingly well. The lesson is not that judgement is useless, but that consistent application of a defined rule is hard to beat.
Designing questions that work
Good questions are observable from public evidence, mutually exclusive in their answers, and tied to a decision. Weak questions are subjective ("is this a good company?"), compound ("is it profitable and growing?") or unobservable ("is the owner willing to sell?"). Rewrite weak questions as observable proxies, or leave them for the conversation.
Order matters. Put identity and eligibility questions first, because failing them makes the rest of the work pointless, and put expensive questions last. Studies of structured decision processes also suggest assessing each dimension separately before forming an overall view, to prevent early impressions from colouring later judgements.
Calibrating and maintaining it
Test the rubric on a sample of companies that the team already understands. Check that reviewers give the same answers, that no question is always unknown, and that the pass rate is plausible. If every company passes, the rubric is not discriminating; if none does, it may be asking for facts that are not public.
Version the rubric. Mandate changes should create a new version, with a record of what changed and why, and completed screens should state the version they used. Re-screen when the version changes, and keep the earlier results so that differences can be explained. This keeps the process disciplined as the mandate evolves.
Related terms
- Disqualifier — A defined condition that ends a candidate's eligibility regardless of favourable answers elsewhere.
- Inter-rater reliability — The consistency of assessments made by different reviewers applying a defined task and decision standard.
- Investment thesis — A proposed explanation of why an acquisition strategy may be worth investigating, expressed as testable assumptions.
Further reading
- Write criteria before collecting answers
- Screening criteria guide
- Disqualifier
- Inter-rater reliability
- Keyword search versus criteria screening
- Article: re-screening when criteria change
- Article: how often to refresh target research
Sources and further reading
- Clinical versus mechanical prediction: a meta-analysis, Psychological Assessment (2000). 136 studies. Mechanical combination of cues equals or beats expert holistic judgement roughly nine times out of ten.
- The robust beauty of improper linear models in decision making, American Psychologist (1979). Even crude equal-weight models outperform expert intuition, because consistency is worth more than precision in the weights.
- Noise: How to Overcome the High, Hidden Cost of Inconsistent Decision Making, Harvard Business Review (2016). Unwanted variability between judges of the same case, and the structured protocols that reduce it.
- A Structured Approach to Strategic Decisions, MIT Sloan Management Review (2019). The mediating assessments protocol: score the parts independently, hold the overall judgement until the end.
To apply these distinctions to your own acquisition criteria, discuss your research brief or inspect the demo platform.