Ask which city the address sits in, because Austin is often the wrong answer
The first call should establish jurisdiction, not just a name and a phone number, since the permit path and the promises you can make change with it.
Mailing addresses hide real boundaries. A house with an Austin address may sit inside West Lake Hills or Bee Cave, or in unincorporated Travis County, and further out you are in Williamson or Hays county rules entirely.
Add jurisdiction as a required CRM field on the first call. It routes the job, sets the permit expectation and tells your estimator what not to promise on the porch.
The same field handles the utility question. Program and rebate eligibility depends on which utility serves that address, and Austin Energy being municipally owned means most statewide Texas advice does not apply inside the city. Tell the caller you will confirm rather than guessing.
Capture the age of the house and the opening count in the same two minutes. Three fields sort a stuck patio door from a sixteen opening job before anyone commits a morning.
The four west-facing openings are the job that closes this month
A whole-house quote a homeowner cannot fund becomes a decision postponed, while a phased quote becomes work on the calendar.
Plenty of buyers here want every window replaced and can pay for half. If your only proposal is all or nothing, you are inviting a competitor to make the smaller offer.
Quote the whole house and then break out a first phase built around the worst heat and the worst operation, usually the west and south elevations. The logic is obvious to the homeowner and it makes the decision smaller.
Record which phase was sold and keep the remainder in the pipeline with a real follow-up date. Phase two is the cheapest job you will ever sell, because the trust is already paid for.
Do not discount to close a phase. The point is to reduce scope, not price, and blurring the two teaches customers to wait for a deal.
Leave the driveway with the number delivered, and count how often that happens
The single largest measurable gap in most window shops is between the appointment and the moment the customer sees a price.
A quote produced on site while the conversation is fresh gets decided against your presentation. A quote emailed three days later gets decided against two competitors' presentations.
Track quote delivery time as a metric of its own, per estimator. Most shops find an average they did not expect and a wide spread between people.
If on-site quoting is not realistic for complex jobs, commit to a stated window and hit it every time. Predictability recovers most of the advantage.
Watch what happens to close rate as the gap shrinks. It is one of the few changes in this trade that moves a number visibly without new spending.
In Mueller and Circle C the architectural committee is a stage in your pipeline
Exterior changes in master-planned neighborhoods often need association review, and a quote that ignores it goes quiet for reasons nobody logs.
Ask early whether the property has an association with architectural guidelines. In many Austin master-planned areas the answer is yes and the homeowner has not thought about it.
Add an approval stage to the CRM with its own follow-up cadence. Without it, the job disappears into a general pending bucket and nobody calls until it is dead.
Help without promising. Provide product documentation and photos the homeowner can submit, and say plainly that approval is the association's decision and worth confirming with them directly.
The same discipline applies where a change to an opening might involve the city. Tree protection and impervious cover rules here are unusually strict, so log it as a stage and tell the customer to confirm requirements with the building department.
Most cancellations happen while the units are on order, not on the porch
The quiet weeks between deposit and install are where a signed job turns back into a refund request.
A homeowner who has paid a deposit and heard nothing for weeks starts to wonder whether they chose wrong. Silence, not price, is usually what unwinds the sale.
Build a fixed contact sequence: confirmation of what was ordered, a production update, a delivery notice, and an install day briefing that says what to move and what to expect.
Measure the cancellation rate between deposit and install as a separate number. Most shops never isolate it and therefore never work on it.
The same sequence produces reviews. A customer contacted four times during the wait is far more likely to write the neighborhood-specific review that helps you in Round Rock or Buda later.
Summer is the only stretch with enough volume to read a test, so plan it in April
One metro will not power small split tests, so save your real experiments for the season that actually produces a sample.
Statistical significance on a button color needs traffic an Austin dealer does not have. Pretending otherwise produces confident conclusions from noise.
Test large things sequentially instead: the intake script, the quote format, the phased offer, the follow-up sequence. Change one, run it for a full month of comparable demand, and read it in booked measures and signed contracts.
Use the heat-driven peak as your test window because the volume is there, and plan what you are testing before it starts rather than improvising in July.
Outside the peak, stop testing and fix defects. Missed calls, unlogged inquiries, quotes never followed up. Defect repair is not glamorous and it is where the money is.
Questions we actually get
- What should we fix first?
- Whatever is unmeasured. In most window shops that means unanswered calls, inquiries that never reach the CRM, and quotes with no follow-up date. Instrument the path from inquiry to measure to signed contract before changing anything, because a month of clean data usually points at a leak nobody suspected and saves you from optimizing the wrong stage.
- Should we count leads, appointments or contracts?
- Contracts, with booked and held measures as the leading indicators. Count openings sold as well as jobs sold, since a four opening job and a twenty opening job look identical in a lead report. Close rate alone hides the problem because it averages several very different kinds of work into one misleading number.
- How long should follow-up continue on an unsold quote?
- Longer than most shops manage. Homeowners here often decide over weeks or months, especially newcomers gathering several bids, and phased work reopens naturally. Build a cadence that continues past the point most competitors stop, keep it useful rather than pushy, and log the reason whenever a quote is finally closed out.
- Is a call answering service worth it during the summer peak?
- It can be, provided the service uses your qualifying questions and passes the address and opening count through cleanly. Test it against your own answer rate before committing. A generic script that collects a name and nothing else usually creates callbacks nobody wants rather than fixing the coverage problem.
- Can we A/B test our landing pages?
- Rarely with confidence at one metro's volume, since small differences need traffic you do not have. Sequential testing of substantial changes, read over a full month against booked measures, is the honest approach. Treat call recordings and lost-reason notes as your real evidence, because thirty conversations will tell you more than an underpowered split test.
