The unanswered call is the largest leak in a window replacement shop
In a trade where the buyer is calling three companies in one sitting, whoever picks up first usually gets the measure.
Pull the call logs and count missed and abandoned calls by hour. In most shops the pattern is obvious and painful: calls go unanswered exactly when the crews and the owner are on a job.
Fix the coverage before touching the website. An answering service briefed on your trade, a shared line that rings a second person, or a scheduler who books measures directly all beat another round of headline testing.
Return every missed call with a text within minutes. A short message naming your company and offering two appointment windows recovers a meaningful share of the people who have already dialled someone else.
Track it as a metric with an owner. Answer rate belongs on the same report as cost per lead, because the two numbers multiply.
Speed to the in-home measure beats every headline you could test
The interval between enquiry and someone standing in the house is the strongest lever in this business, and it is an operations decision rather than a design one.
Book the measure on the first call wherever possible. Offering to call back to schedule adds a step, and every step loses people.
Hold slots for same-week appointments rather than filling the calendar first-come. A booked calendar that turns away fresh demand is a calendar quietly working against the ad spend.
Confirm by text the day before with the technician's name. No-shows on a measure are usually a memory problem, not a commitment problem.
If the homeowner cannot be visited quickly, give them something in the meantime: the wind rating explanation, a photo set of a comparable job, the permit walkthrough. Silence is where the competitor's rep does their work.
Instrument the path from form to measure to signed contract, or you are optimising blind
Lead counts tell you almost nothing in a category where one signed whole-home job outweighs a month of enquiries.
Give every channel its own tracked number, and record calls where permitted so intake quality is visible rather than assumed.
Build CRM stages that reflect the real path: enquiry, contacted, measure booked, measure completed, quoted, signed, installed. Every optimisation argument you will have needs those five middle stages.
Feed the outcome back into the ad platforms where you can, so bidding is trained on booked measures rather than on form fills.
Then report on the ratios, not the totals. Enquiry to measure, measure to quote, quote to signed. A drop in any one of them points at a specific team, week or channel.
The address field qualifies the job before anyone burns a morning driving to it
One well-designed intake question separates a real whole-home job from a visit that was never going to pay for itself.
Ask for the property address at intake and use it. The address settles jurisdiction, since the city, Wellington, Riviera Beach and unincorporated Palm Beach County each administer their own permitting, and it tells the rep which conversation to prepare for.
It also sets expectations on the product. Palm Beach County's Figure 1609.3(1) Ultimate Design Wind Speed map for Risk Category II shows more than one contour across the county, so the design pressure a unit must meet is generally looked up by parcel. A coastal address and one out toward Loxahatchee may not need the same product.
Ask ownership and opening count too. Renters, HOA-restricted buildings and single-opening repairs can be routed rather than sent a full measure appointment.
Qualifying is not the same as gatekeeping. Keep it to a handful of questions, deliver it in a friendly voice, and route the poor fits somewhere useful rather than rejecting them.
One metro will not power small split tests, so test the big things sequentially
A single dealer in one county rarely accumulates enough enquiries to resolve a small difference, which changes what is worth testing at all.
Skip button colours and micro-copy experiments. At realistic volumes you would need a very long run to distinguish a small effect from noise, and the market will have shifted before the test concludes.
Test changes big enough to show up quickly: a form with six fields against one with twelve, an offer of a free in-home measurement against a request for a quote, a page opening with install footage against one opening with a headline.
Run them sequentially over clean, comparable periods, and read them at the booked-measure stage rather than at form submission. A change that raises submissions and lowers appointments is a loss.
Where the traffic genuinely will not support a test, decide with judgement and evidence from the phones instead. Reviewing twenty recorded intake calls teaches more than an underpowered experiment.
Build a follow-up sequence for a purchase homeowners take weeks to decide on
Most quoted window jobs are not lost, they are unfinished, and the dealer who keeps a useful conversation going is the one who eventually installs.
Sequence the follow-up around objections rather than around nagging. Price and financing, the disruption question, the permit and inspection timeline, the warranty on labour.
Send the permit explanation deliberately. The City of West Palm Beach Building Division states that a permit is required to replace windows and doors and that products must be installed per the manufacturer's product approval documents, which reframes the permit from a padded line item into evidence you do the job properly. Point them to the building official for anything specific to their address.
Set honest lead time expectations on ordered product. A homeowner who understands the wait is far less likely to cancel than one who feels forgotten.
Keep contacting quoted-but-unsigned homeowners on a slow cadence after the sequence ends. Whole-home replacement decisions frequently reactivate months later, and the shop still holding the relationship wins that call.
Questions we actually get
- Our leads are low quality. Is that a conversion problem or a traffic problem?
- Usually both, and the intake data tells you which dominates. If most enquiries are repairs, screens or renters, the targeting and the negatives need work. If the enquiries are the right kind of homeowner but few reach a booked measure, the leak is in answering speed and scheduling rather than in the ads.
- How fast do we really need to call back?
- As close to immediately as your operation allows during business hours, and within the same evening for late submissions. The homeowner is contacting several companies in one sitting, so the practical goal is being the first real voice they speak to rather than hitting any particular published benchmark.
- Should we make the form shorter?
- Shorter, but not empty. Six well-chosen fields, with the property address among them, gathers enough to qualify while keeping the friction low. Long forms asking about budget, glass preference and financing before a conversation has happened tend to lose exactly the serious buyers you want.
- What can we realistically A/B test with our traffic?
- Large structural changes read at the appointment stage: form length, offer wording, hero content, whether the measure can be booked on the page. Small cosmetic tests will not resolve at a single dealer's volume, so run big changes sequentially over comparable periods and read them on booked measures rather than form submissions.
- Should we offer instant online pricing for a whole-home job?
- Rarely, because the price depends on the openings, the condition of the bucks, the finishes and the product rating required at that address. A useful middle ground is publishing ranges with the factors that move them, then booking the measure. Precision is what the appointment is for.
