The approval analysis said repeat buyers are worth 2.7× a first-timer. This is where those people are, how many there are, and which ones are about to move. Built from county deed and sales records, nothing purchased.
Two counties, two different data sources, one question: do they tell you to do the same thing? Mostly yes. Where they disagree is marked, and those are the places not to bet.
The Verdict column, in plain words.
AGREE both markets say the same thing, so act on it.
DIFFER they point the same direction but not to the same spot. Take the overlap, not either extreme.
SIZE ONLY not a finding. Harris County is just bigger than Hillsborough, so its numbers are bigger.
CONFLICT they flatly contradict each other. Trust Tampa and read the note under the table.
| Question | Tampa | Houston | Verdict |
|---|---|---|---|
| Does targeting actually work here | |||
| Best tier vs county average | 1.56× | 1.57× | AGREE |
| Core list vs county average | 1.52× | 1.25× | AGREE |
| Homes on the core list Harris County is simply the bigger county | 23,523 | 88,258 | SIZE ONLY |
| Sellers it produces a year follows from the line above | 1,389 | 5,952 | SIZE ONLY |
| Who to put on the list | |||
| Purchase years to target | 2020–2022 | 2020–2023 | AGREE |
| Single hottest purchase year | 2021 | 2023 | DIFFER |
| Home value floor | $350K | $350K | AGREE |
| Above that floor | flat, stops mattering | keeps climbing | DIFFER |
| Coldest buyers, never mail | bought 2010–16 | bought 2010–15 | AGREE |
| Landlords and absentee owners | 0.90× — worse | 1.35× — better | CONFLICT |
| What the market is doing | |||
| Year the freeze bottomed out | 2024 | 2024 | AGREE |
| Direction in 2025 | up | up 14% | AGREE |
| Who is frozen hardest | 12 years in, −54% | 4 years in, −38% | DIFFER |
| County average move rate | 3.88% | 5.39% | NOT COMPARABLE |
Why the last row says not comparable. Houston's records do not distinguish a real sale from an estate transfer, a divorce or a paperwork correction, so its rates are inflated across the board. Tampa's do, and 56% of its recorded transactions turn out not to be real sales. Compare the two markets on lifts and direction, which are unaffected. Never compare them on the raw percentages.
The headline is the first row. Two counties, a thousand miles apart, different states, different economies, different data sources, and the best target tier beats its own market by 1.56× and 1.57×. The targeting travels. That is the finding worth spending money on.
Where they agree, act. Buy owner-occupied single-family homes worth $350K or more, bought 2020 to 2022. Skip anyone who bought before 2016. Both markets bottomed in 2024 and are recovering, so this is a rising tide rather than a falling one.
Where they conflict, trust Tampa. The landlord row is the only outright contradiction, and it has a known cause: Houston counts LLC and trust paperwork as sales, and landlords generate far more of that than homeowners do. Do not build an investor list. The two DIFFER rows are softer, and the safe move is to take the overlap rather than either extreme.
Move rate in 2025 by the year the current owner bought, single family, arms-length sales only. It rises to a clear peak and falls away on both sides.
Nobody who just bought is about to buy again, and nobody who has been there a decade is leaving. The window is years three to five. In Tampa that means the 2020, 2021 and 2022 buyers, right now.
Comparing the same length of ownership across two eras separates a rate effect from people simply settling in. The result is unambiguous, and it is not where anyone expects it.
| Held | Bought | Rate they got | Moved 2015–19 | Moved 2024–25 | Change |
|---|---|---|---|---|---|
| 2 years | 2024 | ~6.7% | 5.71% | 4.46% | −22% |
| 3 years | 2023 | ~6.8% | 7.13% | 5.56% | −22% |
| 4 years | 2022 | ~5.3% | 7.23% | 4.83% | −33% |
| 5 years | 2021 | ~3.0% | 7.35% | 4.52% | −39% |
| 6 years | 2020 | ~3.1% | 6.93% | 4.01% | −42% |
| 9 years | 2017 | ~4.0% | 7.29% | 3.73% | −49% |
| 11 years | 2015 | ~3.9% | 7.01% | 3.36% | −52% |
| 12 years | 2014 | ~4.2% | 7.24% | 3.34% | −54% |
| 15 years | 2011 | ~4.5% | 5.37% | 3.16% | −41% |
How to read a row. Each row follows people at one stage of ownership, not one group of buyers. Take the four-year row: someone four years into their house used to sell 7.23% of the time in a year, and now sells 4.83% of the time, a third less often. The 2015–19 figure is a different set of people who were four years in back then, so Bought and Rate they got describe only who is at that stage today.
The frozen people are not the ones you would guess. Everyone who bought between roughly 2012 and 2021 is holding a mortgage under 4.5%, and that ten-year band is stuck hardest, down as much as 54%. The 2023 and 2024 buyers who paid nearly 7% are barely affected at all, because they have nothing to give up by moving.
That is also what creates the peak on the previous chart. Lock-in did not generate demand at three to five years. It removed everything on either side of it.
Cheap houses move least. Above $350K the difference largely stops mattering, so the useful rule is a floor, not a target.
Landlords are not the prospect. Absentee owners move at 3.40% against 3.88% for people living in the house. That is the opposite of what Houston appeared to show, and the Houston version was an artifact of counting LLC and trust paperwork as sales. Do not build an investor list off this.
Move rate by purchase year and home value, owner-occupied. Darker is hotter. Shading is scaled within Tampa only.
| Value | 2016 | 2017 | 2018 | 2019 | 2020 | 2021 | 2022 | 2023 | 2024 |
|---|---|---|---|---|---|---|---|---|---|
| $750K+ | 4.40 | 4.52 | 4.09 | 4.67 | 5.24 | 5.95 | 5.91 | 5.33 | 3.43 |
| $500–750K | 3.75 | 4.76 | 3.66 | 4.56 | 5.40 | 7.41 | 5.75 | 5.79 | 3.27 |
| $350–500K | 4.02 | 4.20 | 4.64 | 3.75 | 5.68 | 5.94 | 5.80 | 4.36 | 3.35 |
| $200–350K | 3.88 | 4.46 | 4.18 | 4.15 | 4.31 | 4.68 | 4.53 | 3.74 | 4.03 |
| Under $200K | 1.45 | 4.64 | 4.18 | 3.76 | 4.44 | 3.46 | 6.10 | 3.23 | 5.65 |
Every band above $350K peaks in the same place. The cheapest row is greyed because there are too few homes in it to trust year by year.
Tested against the county baseline of 3.88%.
| Definition | Homes | Sold 2025 | Move rate | Lift |
|---|---|---|---|---|
| Owner-occupied, $500K+, bought 2020–22 | 9,869 | 596 | 6.04% | 1.56× |
| Core — owner-occupied, $350K+, bought 2020–22 | 23,523 | 1,389 | 5.90% | 1.52× |
| Owner-occupied, $350K+, bought 2019–22 | 30,678 | 1,684 | 5.49% | 1.41× |
| Absentee, $350K+, bought 2020–22 | 3,295 | 175 | 5.31% | 1.37× |
| Absentee, any value, bought 2019–22 | 14,374 | 513 | 3.57% | 0.92× |
| Owner-occupied, under $350K, bought 2010–16 | 21,230 | 877 | 4.13% | 1.06× |
Houston health warning. Harris County's deed file carries no flag for whether a transaction was a genuine sale. Everything below therefore includes estate transfers, trust moves, divorces and corrections alongside real moves. The rates run high as a result, and they cannot be compared side by side with Tampa's. Comparisons within this market still hold.
This is the finding that survives the data problem, because it is a comparison of the same market against itself. Every length of ownership froze after 2022, and every one of them turned back up in 2025.
| Held | Bought | Moved 2015–19 | Moved 2024–25 | Change | 2024 | 2025 |
|---|---|---|---|---|---|---|
| 3 years | 2023 | 8.71% | 5.90% | −32% | 5.38% | 6.49% |
| 4 years | 2022 | 8.43% | 5.24% | −38% | 4.83% | 5.61% |
| 5 years | 2021 | 7.97% | 5.26% | −34% | 4.95% | 5.53% |
| 6 years | 2020 | 7.75% | 5.10% | −34% | 4.85% | 5.35% |
| 7 years | 2019 | 7.32% | 5.21% | −29% | 4.91% | 5.49% |
| 10 years | 2016 | 6.63% | 4.95% | −25% | 4.99% | 4.91% |
| 15 years | 2011 | 5.20% | 4.01% | −23% | 3.58% | 4.44% |
How to read a row. Each row follows people at one stage of ownership, not one group of buyers. Top row: someone three years into their house used to sell 8.71% of the time in a year, and now sells 5.90% of the time, a third less often. The 2015–19 figure is a different set of people who were three years in back then, so Bought describes only who is at that stage today. The last two columns split the recent figure by year, which is where the recovery shows: 5.38% in 2024, back up to 6.49% in 2025.
Blended, Houston is up 14% off the floor. Deed recording lag would understate 2025 rather than flatter it, so the recovery is at least this large. Tampa shows the same turn.
The same shape Tampa shows, with a steeper slope. Expensive homes turn over far more often than cheap ones.
Houston also appeared to show absentee owners moving at 8.16% against 5.39%. Ignore it. Tampa's clean data shows the reverse, and the gap here is almost certainly LLC and trust transfers being counted as sales. No investor list should be built on this number.
Move rate by purchase year and home value, owner-occupied. Shading is scaled within Houston only, so it cannot be compared to Tampa's grid by colour.
| Value | 2016 | 2017 | 2018 | 2019 | 2020 | 2021 | 2022 | 2023 | 2024 |
|---|---|---|---|---|---|---|---|---|---|
| $750K+ | 6.00 | 5.82 | 6.74 | 5.91 | 5.97 | 7.25 | 7.83 | 10.04 | 12.43 |
| $500–750K | 6.16 | 6.46 | 6.38 | 5.79 | 5.66 | 6.39 | 7.71 | 8.47 | 14.50 |
| $350–500K | 4.61 | 5.12 | 5.17 | 5.23 | 5.42 | 4.98 | 5.99 | 8.78 | 14.87 |
| $200–350K | 3.83 | 3.88 | 4.21 | 4.46 | 4.24 | 4.61 | 5.27 | 6.04 | 14.16 |
| Under $200K | 3.62 | 4.11 | 3.83 | 4.14 | 5.15 | 5.61 | 4.75 | 5.22 | 12.22 |
The 2024 column is greyed. It looks like the hottest column on the board and it is the most contaminated one: a house that has just changed hands generates paperwork, not another move. Tampa's clean data puts one-year move rate at 3.9%, not 14%.
Tested against the county baseline of 5.39%. Read the lifts, not the levels.
| Definition | Homes | Sold 2025 | Move rate | Lift |
|---|---|---|---|---|
| Owner-occupied, $500K+, bought 2022–23 | 19,664 | 1,662 | 8.45% | 1.57× |
| Owner-occupied, $350K+, bought 2022–23 | 41,889 | 3,297 | 7.87% | 1.46× |
| Core — owner-occupied, $350K+, bought 2020–23 | 88,258 | 5,952 | 6.74% | 1.25× |
| Owner-occupied, under $350K, bought 2015–19 | 104,575 | 4,222 | 4.04% | 0.75× |
The anti-list is the reliable part. 104,575 homes under $350K bought between 2015 and 2019 move at three quarters of baseline. That population is larger than the entire target list, and it is close to exactly what a broad Meta campaign optimised on cheap leads will go and find for you.
Houston and Tampa disagree on several things. These four they agree on, which is the only reason to act on them.
The target, stated once: owner-occupied single-family homes worth $350,000 or more, bought between 2020 and 2022. That is 23,523 homes in Tampa producing 1,389 sellers a year, and it sits inside the 88,258-home Houston list producing 5,952. Both run about 1.5× their county baseline.
The rebate is a lock-in offset, and that is the pitch. Trading a 3% mortgage for a 6.5% one on $350K costs about $736 a month. Two percent of a $450K home is $9,000. The rebate covers roughly the first year of the payment difference. Nobody is saying anything to the cohort everyone else has written off as immovable, and the Tampa data shows that cohort is exactly who is stuck.
Tampa's 23,523 homes and Houston's 88,258 are strong custom-audience sizes and good lookalike seeds. Seed the lookalikes off the people who actually sold, not off the whole list.
Tampa's 9,869 top-tier homes at 6.04% and Houston's 19,664 at 8.45%. The Dallas motion already runs at roughly a dollar a card, so this is a known cost against a measured move rate.
The seller pipeline already pulls them. A house on this list hitting the market is a move-up buyer entering the funnel with a 5 to 8% annual probability already established. The purchase year is the prior. The listing is the lead.
It only ever existed in the contaminated Houston numbers. Tampa's clean data has absentee owners moving less than residents, and an absentee list of any value at all comes in below baseline at 0.92×.
Every Houston figure here is inflated by paperwork that is not a sale. Tampa proves how much that distorts, and until Harris can be filtered the same way, Tampa is the market to plan from.