Why Hiring Evidence Should Separate Observation From Interpretation
Most hiring conversations blend three different things without noticing: what an applicant did, what they had in front of them, and what that is assumed to mean. Observation is what happened. Opportunity is what the applicant had available and could have done. Interpretation is what a reader concludes. Calavion keeps them apart on purpose.
Observation: what the record shows
An observation is something the record actually shows: a message sent, a record opened, an action taken, a decision logged, a task left unresolved. It does not require guessing at intent.
Opportunity: what was available
An opportunity is what the applicant had in front of them and could have done at that moment. It matters because the same observation means different things depending on it. “The applicant did not open the payment record” says one thing if the record was available the whole time, and nothing at all if it was not. Opportunity is what turns a bare absence into something a reader can reason about. In a Calavion Replay this is shown directly: an “Evidence available” list, and which records were never checked.
Interpretation: what a reader concludes
An interpretation is a judgment built on top of an observation and an opportunity: that an applicant was careful, decisive, or overwhelmed. Interpretations can be reasonable, but they are conclusions, not facts, and different reviewers can reach different ones from the same evidence.
A worked example (fictional)
Take the fictional cancellation case, where an applicant inherits a handover in which a customer was told an order would be cancelled and refunded, and neither has happened. The records include payment, fulfilment and a warehouse thread. The company, the customer and every record are invented.
- Observation. The Replay shows the applicant wrote to the warehouse and never opened the payment record.
- Opportunity. The payment record was among the records available in the case. The applicant could have opened it at any point.
- Interpretation. One reader concludes the applicant was careless about the refund. Another concludes they judged the warehouse step the more urgent one and meant to return. A third wants to ask.
All three interpretations fit the observation. What separates them is the opportunity, and a question that only the applicant can answer. The record does not say what the applicant was thinking.
A second moment (fictional and scripted)
Take one moment from the scripted Replay for the same fictional case: the applicant, Casey, replies to the customer.
- Observation. The Replay shows Casey’s reply to the customer, with its original wording preserved.
- Opportunity. The payment record, the fulfilment record, the warehouse thread and the policy were all available when Casey replied, and none had been opened.
- Interpretation. Left to the reader. The record shows what was sent and what was in front of Casey. It does not say why.
This is an invented case with a scripted Replay, written to show the three steps. It is not a real applicant, and it is not a result.
What Replay states, and what the reader concludes
Replay reports the order of events and flags, by rule, that a reply was sent before the records were checked. That is a statement about the record. What it means is the employer’s judgment: a reader may see haste, may see confidence in what Casey already knew, or may want to ask. The software does not decide between them.
Why the difference matters
When these blend together, a hiring conversation can end up defending a label (this applicant was not proactive) instead of examining what actually happened. Separating them keeps the record itself stable even when people disagree about what it means.
How this shows up in a Calavion review
Replay reconstructs what happened inside the exercise, moment by moment, without asserting what the applicant was thinking. A hiring manager can look at a specific action, ask the applicant directly about it, and form their own interpretation grounded in something concrete rather than a summary that already did the interpreting for them. Calavion does not give applicants an overall score, generate an ordered shortlist, or recommend whom to hire. The employer orders its shortlist; the software checks that order against approved criteria and asks for a reason where it conflicts.
What this does not solve
Keeping observation, opportunity and interpretation separate does not remove interpretation from hiring; it cannot. Managers still judge. What it does is delay the judgment until after the evidence is on the table, instead of baking a conclusion into how the evidence is reported in the first place.
What we do not know
- Whether separating these three things changes the decisions employers reach. We have no evidence either way.
- Whether every moment in a Replay can be labelled opened or not opened. The example above relies only on records the case defines as available.
Related: How Calavion turns a resolved case into a hiring exercise · The Evidence Ladder · About Calavion