The Takeoff The averages
The average you just read was not a roof like yours
Published figures for this metro sit thousands of dollars apart and every one of them is presented as the answer. They are not lying, mostly. They are averaging houses with nothing in common and letting the reader assume the mean describes theirs. Here is exactly how that happens, and what to measure instead.
The number you came here for exists, but it lives on your roof rather than in a search result. Put the question into a search engine and you will be given confident figures for this metro that sit thousands of dollars apart. Every one of them is presented as the answer. They cannot all be the answer, and understanding why is more useful than picking one.
Failure one: averaging unlike houses
An average is only meaningful when the things being averaged are alike. Roof replacements are not alike. The distribution runs from a small single storey ranch with one shallow simple plane to a large older two storey with a complicated steep roof, two chimneys, plank decking and difficult access, and the second is genuinely several times the job the first is.
Take the mean of that distribution and you get a number that sits in a valley between two humps and describes almost nobody. Publish it as "the average cost in St. Louis" and every reader anchors to it, then feels cheated when their actual roof is priced.
Failure two: measuring different things
One source is quoting material only. Another is material plus labour. A third includes tear off and disposal, which is a substantial fraction of the job. A fourth is quoting a national figure adjusted by a regional cost index rather than anything observed here. A fifth is reporting what people told a survey they paid, which is memory, sometimes years old, sometimes for a different scope entirely.
The same house, five ways of describing its cost
- Material onlyShingles, underlayment, fasteners, deliveredOne number
- Material and labourExcludes tear off and disposalA larger one
- Full replacementTear off, disposal, deck prep, detail, installLarger again
- National average, indexedNever observed in this market at allAnywhere
- Self reported survey figureMemory, unverified, scope unknownUnknowable
Failure three: the calculator is not an estimator
A calculator that asks for your address, roof age and email before producing a range is not primarily a measuring instrument. It is a lead form with an arithmetic animation attached. That is not automatically dishonest, and some of them are built on real aerial measurement, but it is worth being clear about what it can and cannot know.
- It can approximate roof area from imagery reasonably well.
- It infers pitch from shadow, parallax or a property record, and the multiplier it produces is applied to every square foot.
- It cannot see layer count at all.
- It cannot see decking condition at all.
- It cannot see flashing condition at a chimney or a sidewall.
- It cannot see access, where a dumpster fits, or how far bundles must be carried.
So it has guessed most of the inputs, and the output is a range wide enough to be technically defensible and narrow enough to feel like an answer. The number's real function is to be interesting enough that you give up a phone number.
Failure four: the anchor is doing the selling
Once a figure is in a reader's head it does work whether or not it is right. A low published average makes every honest local quote look like an overcharge, which is useful to whoever comes in slightly under it. A high published average makes an ordinary quote feel like a bargain. Neither effect has anything to do with your roof.
So get a measurement instead of an average
This desk prints no number, because any figure published here would carry the four failures above and would do anchoring work on somebody's behalf rather than on yours. What it publishes instead is the geometry, the variable list and the questions. Then go and get the real thing. Keys measures your roof plane by plane and hands you the square count, the measured pitch, the layer count, the decking unit rate, the penetration list and the exact material, in writing, free, before anybody says a price. Six facts about your house beat any average of somebody else's.
What to do with the quotes you have
Stop comparing totals first. Compare inputs first. Line up the square count, the pitch each company measured, the layer count each expects, the decking unit rate, the penetration list and the exact material. If those six agree, the totals are comparable and the difference is about the business. If they disagree, the totals are not comparable at all, and the cheapest one may simply be the one that assumed the most.
Then ask each company how the number changes if the decking is worse than expected, and get that answer in writing. It is the single most predictive question you can ask, because it separates companies that have thought about the unknown from companies that intend to discover it in front of you.
A note on what we do publish
Everything numeric on this site is either geometry you can verify with a calculator, or a public record with the source named on the page: NOAA for hail counts, the National Weather Service for storm ratings, the Missouri Department of Commerce and Insurance for statewide claim totals, the Missouri Revised Statutes for anything legal. If a figure appears here without a source next to it, that is a mistake and we would like to know about it.
Seven inputs, measured
Stop reading averages and get your own seven inputs measured
Every published average is a mean of houses that are not yours. Keys measures the seven that actually decide your figure: square count, pitch per plane, layer count, decking unit rate, penetration list, exact material and the access. Free, about an hour, and the document is yours to take anywhere.
If you would rather talk it through, this is the office line and not a call centre.
(314) 220-2333Keys Roofing, greater St. Louis