An Intervention Candidate is an unvalidated hypothesis about a possible change to systems, data or workflow. It remains outside the sales pipeline. If the client confirms that the issue merits investigation, the supplier can open technical discovery. A commercial opportunity requires a plausible buying process and further qualification.
Separate observation, candidate, discovery and opportunity
Observation: a client has opened a new site or added a maintenance service. This is a fact about the company, subject to source verification.
Intervention Candidate: the change may create a requirement for coordinated asset records or resource scheduling. This is a testable hypothesis, not a confirmed problem.
Sales opportunity: the client has acknowledged a need, is considering a defined change and has a credible route to purchasing work. This is commercially qualified activity, not simply an idea in a researcher’s register.
| Stage | Evidence needed | Owned by | Not yet justified |
|---|---|---|---|
| Observed signal | Identified source, date, relevance and what has actually changed | Researcher | Need, budget or technical solution |
| Intervention Candidate | Business dependency, existing coverage, missing knowledge, falsification test | Researcher / account owner | Project scope, value or probability of closing |
| Qualified for discovery | Customer confirms a substantive issue or improvement goal and agrees to investigate | Account owner + client | Final architecture or procurement decision |
| Commercial opportunity | Buying process, decision participants, approximate scope and commercial next steps | Sales / commercial lead | Guaranteed revenue |
Maintain a separate candidate register
A CRM pipeline is designed to manage commercial transitions. Candidate research is designed to manage uncertainty. Those processes have different denominators, different exit criteria and different stakeholders. A candidate with strong potential significance but very weak evidence should not become a forecastable deal merely to ensure someone follows up.
Keeping candidates separate also makes negative learning visible. A rejected hypothesis can explain that the relevant data already live in another system, that the customer has chosen not to change the process, or that the problem cannot be solved by the provider's services. Those conclusions improve subsequent research and protect account relationships.
Required candidate fields
| Field | Purpose |
|---|---|
| Account and delivered project | Connect the hypothesis to known scope |
| Business dependency | Explain why the issue could matter |
| Observed evidence and source | Separate facts from inference |
| System coverage / unknowns | Identify where the evidence is incomplete |
| Validation question and stop condition | Make the claim falsifiable |
| Status, owner and next review | Prevent research from becoming an unmanaged suggestion list |
Do not attach an arbitrary monetary value to an unvalidated candidate. A large hypothetical project value is not evidence of the client's willingness to act.
Measure validation outcomes, not candidate volume
Count the proportion of candidates reviewed with the account owner, checked with customers, confirmed or rejected, and converted into qualified discovery. Also capture how many added genuinely new information. Pure candidate volume is a poor metric: the system can improve when the research rejects weak hypotheses earlier.
Stage gates: documented evidence creates a candidate; client validation can justify discovery; confirmed buying conditions justify creating a sales opportunity.
For an example, see a hypothetical candidate register. Related reading: white-space analysis without a services checklist.
Proposed analytical method for ProperBizFix. This is a product framework, not an industry standard or an empirically validated probability model.
About these notes. ProperBizFix uses business-model and system-coverage analysis to identify questions worth testing with an existing client. Examples are illustrative unless expressly described otherwise. How the research works.