Your website
Print the real number. It is smaller, and that is the point.
The short answer
ENERGY STAR puts the saving at up to 13 percent, and only when replacing single pane. Publish that number with its conditions, then sell the reason people actually replace windows: the room is unusable in winter.
Somebody has already told your customer that new windows will cut his heating bill by forty percent. Possibly a man at a home show, possibly a flyer, possibly a salesman who sat at his kitchen table for two hours and would not leave. He did not believe it exactly, but it lodged, and now it is the number in his head when he opens your website.
You cannot beat that number honestly, and you should not try to beat it dishonestly. What you can do is the one thing nobody else in your market will: publish the real one.
The real one, and its conditions
ENERGY STAR — a federal program, not a trade body — states that replacing old windows with certified windows lowers household energy bills by an average of up to 13 percent nationwide when replacing single-pane windows. Read that sentence slowly, because every part of it is doing work: up to, average, nationwide, and single-pane.
Single-pane is the crucial one. A house that still has single glazing is the best case in the entire country, and the number quoted is the ceiling of that best case. A customer replacing tired double glazing is somewhere well below it, and he deserves to be told so before he signs, not after his first winter.
Now consider what putting that on your website does. Every competitor in your town is quoting something larger and sourceless. You are quoting something smaller with a government link beside it. The customer cannot verify their number and can verify yours in nine seconds — and the moment he does, every other number he has been shown becomes suspect. You have not lost the argument about savings. You have ended it, in your favor, by being the only one who can be checked.
Then stop selling savings
Because here is the thing the arithmetic will never carry: almost nobody replaces windows to make money. They replace them because there is a room the family stopped using in January. Because the road got busier and the bedroom did not. Because a sash has been painted shut since before they moved in and a fire escape route is not a hypothetical. Because condensation is going through the frames and the repair is now structural.
Those reasons are worth far more than thirteen percent, and none of them needs a calculation. A page organized around them — one section per real reason, with photographs of that exact situation from your own jobs — speaks to the person who is actually about to spend, while your competitors argue about invented percentages with each other.
| Section on the page | What it does |
|---|---|
| The sourced figure, with the link and the conditions spelled out | Ends the numbers argument by being the only checkable claim in the market. Costs you nothing but a smaller headline. |
| One section per real reason — cold room, noise, stuck sash, rot — each with your own photographs | Meets the customer where his decision actually lives. He recognizes his own house and stops comparing on price alone. |
| What you will NOT claim | “We will not tell you these pay for themselves — for most houses they do not, and anybody promising it has not seen your windows.” Nobody else will write that sentence. |
| The survey, described honestly | What you measure, how long you are in the house, and that the number can move afterwards — and by how much, and why. |
| Where the certification sits | Your state’s registration or the relevant scheme, stated plainly — this is what a buyer’s inspector or title company asks for at sale, and most homeowners learn that too late. |
That last line will lose you some jobs, and it is still the right line. A customer told plainly that his cold room is not a window problem does not become a customer — but he becomes the person who says your name when somebody at work mentions windows, which is a longer and cheaper form of the same thing.