A customer has three open estimates. Is that one opportunity or three?
Three, if any one of them could be won while another is lost. Model an opportunity as a distinct decision the customer will make: separate systems, separate properties or separate visits are separate opportunities. Options on the same job, such as good, better and best for one furnace, are one opportunity with variants. Getting this wrong distorts close rate in both directions and makes pipeline value unusable.
The test is independence, not count
Ask whether the customer could plausibly say yes to one and no to another. If yes, they are separate opportunities. If the customer will make exactly one decision and pick a version, it is one opportunity with several priced options.
This single rule resolves most of the arguments. A homeowner replacing a furnace and separately considering a water heater is two. The same homeowner choosing between three furnace tiers is one.
Counting options as opportunities wrecks close rate
Suppose you present three tiers and the customer buys one. Under the wrong model you recorded three opportunities and closed one, which reports as a thirty-three percent close rate on a job you actually won.
Now suppose a competitor presents one option and wins the same job. Their close rate reads one hundred percent. Nothing about performance differs; only the data model does. Any comparison of close rate across teams, locations or time periods is meaningless until everyone counts the same way, which is the same reconciliation problem that shows up whenever revenue reporting spans multiple systems.
Collapsing real opportunities is the opposite error
Merging genuinely independent work into one record hides everything you lost. The customer bought the furnace and declined the water heater, the record closes as won, and the declined work vanishes from the pipeline entirely.
That declined work is often your best follow-up list. It is a known customer, with a known need, at a known price, who has already had a technician in the house. Losing it to a data modeling shortcut is expensive.
Pipeline value is not the sum of open estimates
Adding up open estimates produces a number that is always wrong and always too high. Weight each opportunity by the probability of closing at its current age and stage, using your own historical close rates by job type and estimate age rather than a stage percentage someone typed into a configuration screen years ago.
The weighted number is smaller, less exciting, and actually forecasts. It also degrades gracefully: as an estimate ages past the point where your history says it usually closes, its contribution falls automatically. Building that weighting from real completed-job data is a straightforward piece of business intelligence work once the opportunity records are modeled consistently.
Topics: opportunities · pipeline · close rate · data model
Have a version of this question about your own business?
The useful answer usually depends on which systems you run and how they're connected. That's a conversation, not a blog post.