A seller updates a lead's annual revenue and employee count fields after a productive discovery call, expecting the predictive score to jump immediately so the lead moves up the prioritized work queue. Checking back an hour later, she finds the score unchanged. A consultant reviewing the configuration confirms predictive scoring is enabled and functioning correctly. What should the consultant tell her?
Select an answer to reveal the explanation.
Short Explanation
The scoring engine is not sitting there watching every keystroke the seller makes. It runs on a cycle, so it processes a batch of updated records at set intervals rather than reacting the instant someone hits save. That means an hour of silence after editing a couple of fields is completely normal, not a sign that anything is broken. The revenue and headcount changes she made are the kind of details this kind of model typically cares about, so they are very likely queued up and waiting for the next pass, not being ignored. Nobody needs to delete and rebuild the record, and nobody needs to retrain anything just because the number has not moved yet. The right move is simply to wait for the next scheduled refresh and check again. Setting that expectation up front saves everyone from chasing a problem that does not actually exist, and it keeps trust in the scoring system intact instead of prompting people to second-guess a tool that is working exactly as designed.
Full Explanation
The correct answer is A. Dynamics 365 Sales predictive lead scoring does not recompute a lead's score the instant a field is saved; the model processes records on a periodic refresh cycle, so a change made minutes ago will not be reflected until the next scheduled run completes. The consultant's job here is to set the seller's expectations correctly rather than treat this as a defect. Option B is incorrect because revenue and employee count are exactly the kind of firmographic attributes predictive scoring commonly weighs, so ruling them out as irrelevant is unfounded and would mislead the seller about how the model works. Option C is incorrect because predictive scoring evaluates leads on an ongoing basis as data changes and refresh cycles run, not just once at creation, so recreating the record is unnecessary and would actually lose the lead's history. Option D is incorrect because a delay before a refresh cycle completes is normal, expected behavior, not evidence of a broken or untrained model, and retraining is a disproportionate response to a timing question. Recognizing the refresh cadence keeps the consultant from chasing a non-issue and from prescribing unnecessary rework.