February traffic arrives from a cold bedroom at ten at night
The visitor profile in a real winter market is a homeowner who just felt the draft, on mobile, alone, and unlikely to call until morning.
Design for that person and the rest follows. Large tap targets, a form above the fold on mobile, a phone number that dials on tap, and a clear statement of what happens after they submit.
Because the visit often happens after hours, the site is the only salesperson on duty. An after-hours submission should return an immediate confirmation saying when a human will call, and the callback should happen when promised. Silence overnight is where most of these inquiries die.
Desktop still matters for the comparison stage. Someone reading about insert versus full frame at the kitchen table on a laptop is closer to a decision than the midnight visitor, and they need longer content and clearer specification detail.
Do not build a site that only looks right on the office monitor. Test it on an ordinary phone in a driveway in Bellevue or Donelson, on whatever signal you actually get there.
The first screen has to name your counties before it names your manufacturer
Brand logos mean nothing to a visitor whose first question is whether you drive to them at all.
The hero needs three facts: what you replace, the geography you cover, and the next step. Windows and exterior doors, Davidson and Williamson counties, free in-home measure. Nothing else earns that space.
Manufacturer badges belong lower on the page. They reassure a buyer already comparing specification, and they confuse one who does not yet know whether you serve Hendersonville.
A photo of a house that looks like Middle Tennessee outperforms a stock image every time. A brick ranch, a prewar cottage in Sylvan Park, a newer build with a portico. Recognizable beats aspirational.
Keep the promise honest. If you do not cover Rutherford County past a certain point, say so on the page. Filtering out the drive you will not make is worth more than the inquiry you decline later.
Insert versus full frame is the page that stops the price question
Most of the price objection in this trade comes from a buyer who does not understand what they are comparing between bids.
Build one substantial page explaining both approaches, in plain language, with a diagram. When the existing frame is sound, when it is not, what happens to interior and exterior trim, how the sizes are affected, and why two quotes can differ enormously and both be reasonable.
Add the older-home version. Original wood sashes in a Germantown or 12 South house raise questions about trim, matching, and whether the character survives. Explain how you handle it and show a completed example.
Say plainly that some neighborhoods carry review or overlay considerations and that requirements vary by jurisdiction, so they are worth confirming with the local building department before an order is placed. Do not publish rules or promise an approval.
A page like this is the most linked and most shared asset these sites have, because nobody else in Middle Tennessee writes it and every homeowner needs it.
The quote form should ask what is failing, how many openings, and when they need it done
A five-field form capturing the right five fields books more measures than a two-field form and disqualifies fewer good jobs than a twelve-field one.
Ask for the address, not just the city. The address tells you the county, the drive, the likely house age and often the neighborhood, which is more qualification than any dropdown provides.
Ask the opening count as a range, the rough age of the current windows or door, what is wrong with them, and whether a date is driving the project. A rental turnover, a closing, or a house going on the market changes the whole conversation.
Add an optional photo upload. One picture of the sill and the sash answers more questions than a paragraph, and it lets an estimator arrive prepared rather than surprised.
Skip fields you will not use. Preferred contact time is useful. Budget dropdowns push honest buyers away and invite low answers from people who have not priced anything yet.
A page that loads in a Donelson driveway beats a page that looks good in the office
Speed on an average phone on an average connection converts more inquiries than any layout change you are considering.
The usual culprits are predictable: uncompressed project photos, a video hero, a chat widget, a financing calculator, review widgets, and three tracking scripts nobody audits. Each one is defensible alone. Together they cost you the visitor.
Project photography is your best evidence and your heaviest asset. Keep it, but serve modern formats at the right dimensions and load below-the-fold galleries only when the visitor scrolls to them.
Third party widgets should earn their weight. If the chat tool produces inquiries you would not otherwise get, keep it and load it late. If it mostly answers questions the FAQ already answers, remove it.
Measure on a real device with a throttled connection, not on office broadband. Middle Tennessee has plenty of places where a phone falls back to a weaker signal, and that is where a bloated page loses.
Leave savings math, historic approval promises and stock winter photos off the site
The claims that get a window dealer into trouble are the ones that sound harmless in a headline.
Do not publish a specific energy savings figure, a payback period, or a utility bill reduction. Performance varies by house, product and installation. Say ratings vary by product and that any rebate or incentive eligibility is worth confirming with the program administrator or your own accountant.
Do not promise a historic or overlay approval, and do not state permitting rules for a county. Practice differs across Davidson, Williamson, Rutherford, Sumner and Wilson. Say it is worth confirming with the local building department.
Do not use stock photography of houses you did not touch, and do not retouch a before-and-after. A homeowner comparing three sites can spot both, and one caught exaggeration costs more than the photo ever earned.
Never publish an insurance outcome or imply a claim will be covered. If a window failure ties into a claim, say the answer depends on the policy and is worth confirming with their insurer or their own counsel.
Questions we actually get
- Do we need a whole new site or can we fix the one we have?
- Usually a rebuild of the pages that matter beats a full replatform. If the site loads reasonably, runs on a maintainable system and the content is merely wrong, rewriting the homepage, the product pages, the county pages and the form gets most of the result for a fraction of the cost. Replatform when the site is slow at the foundation or you cannot edit it without a developer.
- Should we publish prices?
- Publish the mechanics of pricing rather than a price. Explain what drives the number: opening count, frame material, glass package, whether an insert is possible, custom sizing and lead time. If you are comfortable posting a range you will honor, do it and date it. A number you quietly abandon at the kitchen table costs more trust than it bought.
- Is a financing calculator worth the page weight?
- Only if financing closes a meaningful share of your jobs. If it does, load it on its own page rather than on the homepage, so visitors who need it get it and everyone else gets a faster site. If it exists mainly because a lender offered the widget, remove it and mention financing availability in text.
- How many pages does a dealer like us actually need?
- Fewer than most agencies propose. A homepage, a page per product line you genuinely sell, the insert versus full frame explainer, a page per county you cover, a real project gallery, an about page with faces and a contact page with a working form. Add submarket pages only where you have work to show in Franklin, Brentwood, Murfreesboro or Hendersonville.
- Should the site have live chat?
- Test it rather than assume. Chat helps when a real person answers within a minute during business hours and when the after-hours version collects a phone number instead of pretending to be staffed. Chat hurts when it is a slow script that blocks the page and answers questions your FAQ already covers. Judge it on booked measures, not on chat volume.
