Home services and trades
Lost quotes and estimates: why the jobs you did not win matter
By SourceX Editorial · Updated
Short answer
Lost quotes and estimates matter because they hold the other half of every sales decision. Sold jobs show what customers bought; unsold estimates show what they turned down, at what price and after which follow-up. Keep the estimate ID, options and prices presented, dates, presenter, lead source, a coded lost reason, any competitor named and the follow-up history.
Key takeaways
- Wins alone show what sold; wins and losses together show where price, options and timing changed the customer's decision.
- A coded lost reason is worth more than a long note, because it can be counted and compared.
- Unsold estimates often disappear in migrations and cleanups, so protect them on purpose.
- Quoting and sales tools need both outcomes, because a model cannot learn where a decision tips from one side of it.
Why do lost estimates matter?#
Lost estimates matter because they record the decisions customers made against you, which sold jobs cannot show. A list of sold replacements tells you what people bought. Only the unsold estimates show where the price was too high, which option tier was ignored and which follow-ups never got a reply.
Most trades businesses already create this record. Every comfort advisor's proposal, every repair-versus-replace option and every commercial bid exists somewhere in the field service platform or a sales tool. The problem is that losses are rarely kept as carefully as wins.
What lost estimates reveal that sold jobs cannot#
Lost estimates reveal the edges of your pricing and sales process. The comparison shows which questions sold jobs can answer on their own and which need both outcomes.
Use the comparison in regular pricing and sales reviews, not just once. A replacement option that keeps losing on price may need a different tier, not a lower price, and a lead source that produces many visits but few sales may need a phone screen before an advisor drives out. Be careful with small samples: a few lost estimates in a slow month say little on their own.
| Question | Sold jobs only | Sold and lost together |
|---|---|---|
| Which option tier do customers choose? | Which tiers sold | Which tiers were offered and passed over |
| Where does price stop working? | Prices that were accepted | Prices that were declined, by job type |
| Does follow-up change outcomes? | Follow-ups on deals that closed | Close rates with and without follow-up |
| Which lead sources justify an in-home visit? | Revenue by source | Visits that produced nothing, by source |
| How do advisors compare? | Each advisor's sales | Each advisor's close rate on similar leads |
| When do customers repair instead of replace? | Replacements sold | Replacement estimates that ended in a repair |
Fields to keep on every lost estimate#
Every lost estimate should keep enough fields to rebuild the decision later. Most are captured already; the usual gaps are the reason, the follow-up and the final status.
Make the reason and the final status required before an estimate can be closed, and keep the field list the same across technicians, advisors and estimators. Consistent fields matter more than complete notes, because they let you compare one season, one advisor or one job type against another.
- Estimate ID, linked customer, location and the originating call or job.
- Date presented and date marked lost or expired.
- Every option presented, with line items and prices.
- Financing offered, and whether it was accepted or declined.
- Who presented: technician, comfort advisor or estimator.
- Lead source and job type.
- Equipment age and condition from the inspection.
- Lost reason as a code, plus an optional short note.
- Competitor named by the customer, if any.
- Follow-up attempts with dates, channel and response.
- Final status: lost, repaired instead, deferred or no decision.
How to standardize lost reasons#
Standard lost-reason codes turn a pile of notes into something you can count. Keep the list short enough that advisors use it consistently, and agree on definitions with the sales team before rollout.
Avoid a catch-all code. If one code ends up covering most losses, the list is not doing its job, and the sales manager should split it into reasons the team can act on.
| Code | Use when | Common mistake |
|---|---|---|
| Price | The customer said the total or monthly payment was too high | Using it when the real reason was never asked |
| Chose repair | The customer repaired the existing system instead | Leaving the estimate open after the repair job |
| Chose competitor | The customer hired another contractor | Recording a competitor the customer never named |
| Financing declined | The customer did not qualify or declined the terms | Storing credit details instead of a simple flag |
| Timing | The customer deferred to a later season | Marking it lost instead of deferred |
| No response | Follow-ups got no reply | Using it before the follow-up plan is finished |
| Not a fit | Outside the service area or scope | Mixing it in with price losses |
Where lost estimates disappear#
Lost estimates disappear through ordinary housekeeping. Many platforms let users dismiss or expire unsold estimates, some teams purge old ones to tidy dashboards, and software migrations often import only sold or open estimates.
Advisors also keep proposals in sales apps, on tablets or as PDFs in email that never sync back to the job record. Before the next software change or cleanup, decide that unsold estimates are kept with their IDs, and name them explicitly in any export plan.
Why AI developers value win-and-loss records#
AI developers value win-and-loss records because quoting, pricing and sales-assist tools have to learn where a decision tips, and that boundary only exists when both outcomes are present. A dataset of accepted quotes cannot teach a model which offers fail.
Linked estimate histories also show sequences: inspection findings, the options built from them, the conversation, the follow-up and the result. Customer names, addresses, phone numbers and any financing or credit information are removed before such records leave the company, while the outcome codes stay.
Illustrative: a replacement team starts coding its losses#
Illustrative: a fictional residential HVAC company runs comfort advisors for replacement sales and presents good, better and best options on tablets. Unsold estimates expire automatically, and lost reasons, when recorded at all, are free-text notes.
The sales manager turns off automatic purging, adds a coded lost-reason field and a required final status, and asks advisors to log follow-ups in the platform rather than on their phones. The office backfills codes for older estimates from notes and marks them as backfilled.
After a few selling seasons, the team can see which option tier it loses at and which lead sources rarely close. The linked win-and-loss history also becomes one of the record families the owner lists in the company's data inventory.
How SourceX looks at estimate histories#
SourceX assesses estimate histories with the SourceX Enterprise Data Value Framework, a SourceX-developed methodology that gives qualitative ratings rather than prices. Several of its drivers apply directly: human-generated signal and domain expertise in how advisors build and present options, uniqueness of a company's own pricing decisions, and AI utility for quoting and sales tools. Privacy burden and preparation cost reduce net value, which is one reason credit and financing details stay out.
If the company proceeds, the records move through the SourceX five-step transaction: Supply, Rights, Preparation, Approval and Delivery. The company keeps ownership and approves every step.
Frequently asked questions
How long should we keep lost estimates?
Treat lost estimates like the rest of your job history and set the period in a written retention policy. Keep the decision fields and outcome even when you minimize personal details. Ask counsel or your accountant whether any records must be kept for a set time in your state.
Should advisors record which competitor won?
Only when the customer volunteers it, and in neutral terms. A competitor field helps spot patterns, but notes about competitors should stay factual. Avoid guesses or opinions that would read poorly if anyone outside the company ever saw them.
What about financing applications on lost deals?
Keep a simple flag showing financing was offered, accepted or declined. Credit applications, scores and lender decisions belong with the lender and should not sit in the estimate record. They are excluded from any data package.
Do lost commercial bids count?
Yes. Lost commercial bids, with scope, price and the reason given, can be some of the most informative records a contractor has. Check the bid documents first, since plans and specifications from the owner or general contractor may carry confidentiality terms that limit reuse.
Can we recover lost reasons for old estimates?
Partly. Notes, emails and follow-up records often show the reason, and the office can backfill codes from them. Mark backfilled codes as such, and leave the field blank when the reason is unclear rather than guessing.
Related resources
- QuestionCan SaaS data be licensed?
- QuestionCan CRM data be licensed?
- InsightCan you license CAD and engineering drawings to AI companies?
- InsightCan you license code reviews and pull requests to AI companies?
- InsightCode snapshot vs full git history: what to include in a code license
- SolutionEnterprise data: the records of how organizations actually work
See if your company qualifies
A short company assessment. No data uploads are needed.