Bid Pipelines in HubSpot: The Stages a Contractor's Pipeline Actually Needs

A contractor reports a 22% win rate to a bank, and nobody in the business believes it. Two of last quarter's largest jobs were awarded and never started because the developer could not close financing, one was cancelled before any award was made, and a fourth was withdrawn when the schedule moved. All four sit in a stage called Closed Lost beside the bids that were genuinely beaten on price, and the number that comes out of that is arithmetic performed on four different events.

This article sets out how to model a bid pipeline in HubSpot so the win rate survives contact with how contracting actually works. It begins with the difference between a bid and a sale, then names the four ways a bid ends, then gives a stage model and the fields that hold the distinction. It closes with the reporting that becomes possible once award and win are separate facts, and what this model does not solve.

Three terms are used precisely. A bid is a priced submission against a scope somebody else defined. An award is the client selecting that submission. A win is work that actually proceeds to execution, which is a different event on a different date, and frequently does not follow from an award at all.

The Shape of a Bid, Against the Shape of a Sale

A standard sales pipeline assumes the seller controls the pace. Discovery happens when the seller books it, a proposal goes out when the seller writes it, and the deal closes when the buyer signs, so the stages describe a process the seller is running.

A bid inverts that. The invitation arrives on someone else's schedule, against a scope and a specification the contractor did not write, with a submission deadline that is fixed and a decision date that routinely is not. The contractor controls the price and the quality of the submission and almost nothing else, which means a stage representing seller activity describes very little of what is happening.

The practical consequence is that the useful stages are states of the bid rather than steps in a process, and HubSpot's deal pipeline is perfectly capable of holding either as long as somebody decides which one it is modelling (HubSpot pipelines).

Four Terminal States, Not One

The defect that produces an unbelievable win rate is a pipeline with a single losing outcome, so here are the four that a contractor's work actually ends in.

Not submitted. An invitation arrived and the contractor declined to bid, because the scope was wrong, the schedule was impossible or the general contractor was one they will not work with. This is a decision worth recording against the invitation, and it is emphatically not a loss.

Lost at award. The bid was submitted and somebody else was selected. This is the only one of the four outcomes that belongs in a win rate as the term is conventionally understood.

Awarded, not proceeded. The contractor was selected and the job never started, because financing failed, permits did not come through, or the client shelved the project. The work was won commercially and produced no revenue whatsoever.

Cancelled before award. The project itself went away while the bid was outstanding, so no decision about the contractor was ever made.

Collapsing those four into Closed Lost destroys the distinction between losing on price and losing to a condition nobody controlled, and that distinction is the whole content of a bid review.

The Stage Model

Seven stages, and the argument for each is that a bid genuinely sits in it for a measurable period rather than passing through it instantly.

Invitation received. The bid exists as a record from the moment it arrives, including the ones that will be declined, because a pipeline that only contains pursued work cannot report on what was turned down or why.

Bid or no-bid decision. A short stage with a real decision at the end of it. Recording the decline here rather than deleting the record is what makes capacity and scope-fit visible later.

Estimating. The longest stage in most contractors, and the one consuming the resource that actually constrains the business, which is estimator time rather than salesperson time.

Submitted. The bid is out and the contractor is waiting. Anything that happens now is on somebody else's schedule.

Awarded. Selected, not started. This is the stage most pipelines lack, and its absence is what forces an award to be recorded as a win on a date the work had not begun.

In execution. The job started. This is the win, and it is where a deal becomes commercial history rather than a forecast.

Closed. Terminal, with the reason held in a constrained field rather than encoded in the stage itself.

Award and Win as Separate Fields

The stage says where a bid is now, and two dated fields say what eventually happened to it, and the reporting depends on the fields rather than the stage (HubSpot properties).

award_date records when the contractor was selected. execution_start_date records when work actually began. A bid with the first and not the second, thirty days on, is the single most useful thing a contractor's pipeline can surface, because it is revenue that was won and is not happening.

A third field carries the terminal reason, constrained to the four states above plus a price loss, so a report can separate them. Free text here produces sixty spellings of "financing" and no report at all.

Line items hold the trade breakdown where a bid is priced by section, which keeps the estimate's structure available to reporting instead of flattening it into a single amount (HubSpot line items).

Reporting That Becomes Possible

Four numbers, none of which can be computed from a pipeline with one losing stage.

A bid-to-award rate measured only against submitted bids, which is what a contractor is actually competing on and the number an estimator can influence.

An award-to-start rate, which measures the quality of the client rather than the quality of the bid, and which identifies general contractors whose awards routinely evaporate.

A no-bid rate with reasons, which says whether the business is being invited to the wrong work, and is the earliest available signal that a market has shifted.

Estimating hours against awarded value, which is the closest a contractor gets to a cost of sale, and which a pipeline holding declined bids can compute and one that deletes them cannot (HubSpot custom reports).

The Business Case

What this buys is a win rate that can be shown to a bank, a surety or a board without being explained away, and an early view of awarded work that is not starting.

What it costs is discipline at two moments: recording an invitation that will be declined, and setting a date and reason on a bid that ends. Both are seconds of work and both are routinely skipped, because the person closing the record gets nothing from the field and the business gets everything.

The case is strongest where bids are large, few and slow, so that each misclassification moves the reported rate materially. It is weakest in high-volume residential work, where the distinction between award and start collapses because the two happen within days of each other, and the extra stage earns nothing.

Boundaries of This Model

Nothing here sets bonding capacity, surety requirements or how a contractor should price risk, which are determinations for the business and its broker rather than for a pipeline.

The model assumes bids are discrete and separately awarded. Contractors working under a master agreement with releases against it have a different object problem, where the agreement is the relationship and each release is a job, and flattening the two into one deal loses the level at which the commercial relationship lives.

It also assumes the contractor learns the outcome. A meaningful share of submitted bids simply go quiet, and a pipeline needs a dormancy rule for those rather than a fifth terminal state invented to hold them.

In Summary

A bid ends in four different ways and a standard sales pipeline offers exactly one of them, which leaves a contractor reporting a win rate that describes nothing anybody recognises.

Model the states rather than the seller's activity: invitation, bid or no-bid, estimating, submitted, awarded, in execution, closed. Keep award and execution as separate dated fields, and constrain the terminal reason so it can be reported on.

Awarded work that has not started after thirty days is the number this model exists to produce. It is revenue the business has already counted and is not receiving, and no pipeline with a single Closed Lost stage can see it at all.