What overpricing actually costs
Every agent tells you not to overprice. Almost none of them can show you the number. Here it is, from every listing that ended in central Illinois since 2019 — the ones that sold and the ones that gave up.
- Homes that sold asked
- +9%
- above modelled value
- Homes that did NOT sell asked
- +16.9%
- above modelled value
- The difference
- 8 pts
- 18,681 sales, 2,927 failures
The part most people get wrong
Sellers who succeeded did not price at market value. They asked roughly +9% above it — room to negotiate is normal and it works. Asking above value is not the mistake.
The homes that never sold asked about +16.9% above. That is the whole difference between the two groups: not whether you ask above value, but how far.
County by county
The pattern holds everywhere we could test it. Shelby County is the most unforgiving of the five.
| County | Sold, asked | Unsold, asked | Difference | Days to sell | Days before giving up |
|---|---|---|---|---|---|
| Macon10,269 / 1,486 | +8.9% | +15.9% | 7 pts | 7 | 70 |
| Coles3,866 / 636 | +9.6% | +16.3% | 6.7 pts | 7 | 78 |
| Effingham2,302 / 418 | +6.6% | +14.9% | 8.3 pts | 11 | 103 |
| Shelby1,381 / 241 | +11.4% | +29.6% | 18.2 pts | 11 | 164 |
| Moultrie863 / 146 | +9.4% | +15.4% | 6 pts | 13 | 107 |
Counts under each county name are sales / unsold listings behind that row.
The time cost is worse than the price cost
Look at the last two columns. Homes that sold did it in 7–13 days. Homes that failed sat for 70–164 days before their sellers gave up — and most of them then relisted lower, having lost a season and the interest of every buyer who had already seen it sitting.
So what should you ask?
That depends on your house, not on a county median — its condition, its location, and what genuinely comparable homes nearby have sold for recently. That is the conversation we would rather have with you before the sign goes up than three months after it.
Get a valuationHow these numbers were worked out
A statistical model was fitted to closed sales only — never to the listings that failed. It learns what the market pays for size, land, age, location, a garage, a basement, central air, and the condition the listing describes.
Every listing was then scored against that model and compared with what it was originally asking. Homes that sold are the control group: if successful sellers also asked well above modelled value, the failures’ figure would mean nothing on its own. It is the difference between the two that carries information.
Price per square foot was tried first and gave a misleading answer, because homes that fail to sell are not a random sample — they are older, larger and more rural on average, and a per-square-foot comparison holds none of that constant.
What this does not say: that overpricing causes a home to fail. A house can also struggle because it is unusual, poorly presented, or in a thin market, and the model cannot see any of that. The model itself carries a median error of about 15%, so no individual home’s figure would mean anything — only the comparison between two groups of hundreds does.