Why Unsold Estimates Go Cold
A replacement proposal has a shelf life. Nobody in the business owns it, no system chases it, and somewhere around week three it stops being a live opportunity and becomes a number in a report.
Unsold estimates are the largest recoverable revenue leak in most HVAC businesses — larger than pricing, larger than efficiency, larger than almost anything on the operations side.
They go cold for one structural reason: nothing in the business is responsible for them after the technician drives away. Fixing that is a process problem with a software component, in that order.
The size of the leak
Run your own version of this calculation. It takes two minutes and it usually changes the conversation.
| Example shop | Yours | |
|---|---|---|
| Replacement estimates per month | 40 | ____ |
| Average value | $9,000 | ____ |
| Total quoted per month | $360,000 | ____ |
| Close rate | 30% | ____ |
| Closed | $108,000 | ____ |
| Left unsold, monthly | $252,000 | ____ |
| Recovering just 5% of that | $12,600/month | ____ |
The last row is the point. You are not trying to close everything — most of those customers were never going to buy from you. Recovering a single-digit percentage of the backlog is worth more than most operational improvements available to a shop of this size, and it requires no additional leads, marketing spend or headcount.
Why they go cold
Four causes, in order of how much damage they do.
Nobody owns it
The technician presented the proposal and moved to the next job. The office assumes the customer will call. The customer assumes the contractor will follow up. Both wait. The estimate expires by default rather than by decision.
There is no sequence, only intention
Most shops genuinely mean to follow up. In practice it happens when someone remembers, which correlates with how quiet the week is rather than how valuable the estimate was. A busy July means the highest-value proposals of the year get the least attention.
The follow-up says nothing
“Just checking in to see if you had any questions about that estimate” gives the customer nothing to respond to. It is a nudge, not information, and it can be ignored without cost. Three of those and the customer has learned to ignore you.
The reason for delay was never captured
A homeowner hesitating over $9,000 has a specific reason: they are getting other quotes, waiting on a spouse, hoping to nurse it through one more season, or the number is beyond what they can arrange. Each of those needs a different response. Almost nobody records which one it is, so the follow-up is generic by necessity.
How the window actually closes
The sequence that works
Four touches, each with a different job. What matters is not the cadence but that every message contains something the customer did not already have.
Day 2 — Answer the question they did not ask
Not “did you have any questions”. Instead, address the objection you know is coming: the efficiency difference between the option they are considering and the cheaper one, or what the warranty actually covers, or a note about the rebate that applies.
Send it fast. This is the message with the highest chance of a reply, and every day it waits costs some of that.
Day 7 — Remove the obstacle
By now they have probably decided the number is the problem. So change the number’s shape rather than the number: monthly financing figures, or a genuinely reduced-scope option that solves the immediate problem for less.
Framing matters here. “A smaller system that fits your budget” respects the constraint. “Here’s a cheaper option” implies the first quote was padded.
Day 21 — The cost of waiting
Concrete, not vague. What they will spend on repairs to the old system over another season. What their bills look like at the current efficiency versus the proposed one. If a rebate or price is expiring, when.
This is the message that moves the customer who has been telling themselves the system will last one more year.
Day 60 — Expiry and release
Prices do expire, legitimately, because material costs move. A clear note that the quoted price stands until a specific date is honest and it creates a real deadline rather than a manufactured one.
Then close the estimate and record the reason. It is not a lost opportunity — it is a customer to approach differently, likely when their system fails.
Segmenting: not every estimate deserves four messages
| Estimate | Treatment |
|---|---|
| High value, agreement holder | Full sequence, plus a phone call from the owner or comfort advisor at day 7 |
| High value, new customer | Full sequence. They are almost certainly comparing quotes, so the day 2 message matters most |
| Mid value repair | Day 2 and day 21 only. The economics do not justify four touches |
| Low value | One automated reminder. Anything more costs more than it recovers |
| Explicitly declined | Stop. Record the reason. Move them to a long-term nurture list |
The distinction that pays is the first row. An agreement holder who has just been quoted a replacement is the warmest opportunity in your entire pipeline, and it is worth a human phone call rather than an automated email.
The data nobody captures
When an estimate closes without a sale, record why in a structured field with fixed options rather than a free-text note:
- Price — beyond budget
- Price — chose a competitor
- Timing — deferring to next season
- Decision — waiting on another person
- Scope — wanted something different
- Repaired instead
- No response
Six months of this data tells you things no amount of instinct will. If a third of your losses are “chose a competitor”, you have a pricing or presentation problem. If a third are “no response”, you have a follow-up problem. If a third are “deferring to next season”, you have a seasonal nurture list worth building — and those customers are your best prospects next spring.
Software can send the sequence. It cannot decide what goes in the day 21 message, or notice that an agreement holder deserves a phone call, or capture a reason nobody thought to ask for.
Write the four messages and the segmentation rules on paper first. Automating an undefined process just makes generic follow-up arrive on schedule.
What the software has to do
Four capabilities. Most field service platforms handle the first two and not the last two, which is the gap that separates a customer database from a pipeline — covered in more detail in the comparison of CRM and field service management.
| Capability | Usually present? |
|---|---|
| Store the estimate with its value and date | Yes |
| Show a list of open estimates | Yes |
| Trigger multi-step sequences automatically by age | Often not |
| Branch the sequence by value or customer type | Rarely |
The test to put to any vendor: show me how a $12,000 unsold estimate moves through automated follow-up at day 2, day 21 and day 60, with different messaging at each step. A list of pending estimates is not an answer to that question.
Frequently asked questions
How long should I keep following up on an estimate?
What should the first follow-up actually say?
Is it worth following up on small estimates?
Should follow-up be automated or personal?
Why do I need to record the reason an estimate was lost?
My field service software has an estimates list. Is that enough?
What to do next
- Run the calculation in the table above with your own numbers. Total your unsold estimates from the last 90 days.
- Write the four messages before automating anything. Each one has to contain new information.
- Add a structured “reason lost” field with fixed options. Six months of it will tell you more than a year of guessing.
- Pick out every agreement holder with an open replacement estimate and call them this week. That list is the warmest revenue in your business.
- How flat-rate price books work — pricing the options you present
- Estimating & Invoicing — proposals, pricing and payment
- How to evaluate HVAC software — the scenario that tests this properly
The estimate values, close rates and decay curve in this guide are illustrative examples used to explain a method, not measurements from a specific business or an industry benchmark. Run the calculation with your own figures.