Banking Switch Potential Scores
The likelihood that a household moves its primary checking relationship to a new bank in the next 12 months.
Across 139.6M U.S. households in the Semcasting file, 14.0M clear the high-fit bar for banking switch potential (score 64+, the top 10% nationally). Among metro areas they concentrate in Clarksville, Greeley, Lubbock: in the Clarksville metro area, 44.5% of households are high-fit, 4.43× the national rate. At $1.50 to $3.00 CPM, these audiences are priced at approximately 50% or less of the market rate, with results reported every week.
Fair Lending qualified
No race, sex, age, marital, familial or public-assistance inputs, and no proxies for them.
Household + zip signals
Semcasting household fields combined with Census, HMDA, IRS, FDIC, BLS and CFPB data for the zipcode.
State and zip indexes
Scores roll up to every U.S. state and zipcode, indexed to the national average (100).
Households national audience
- U.S. households scored
- 139.6M
- High-fit (score 64+)
- 14.0M
- Qualified (score 57+)
- 37.0M
- U.S. zipcodes indexing 120+
- 643
Share of households clearing the high-fit bar in the top metro area vs the nation. Qualified tier (score 57+, the top 25%): 37.0M households. Average score 50.6.
States highest high-fit share
| State | High-fit | Share |
|---|---|---|
| District of Columbia | 88K | 25.7% |
| Colorado | 630K | 24.9% |
| Nevada | 301K | 22.9% |
| Georgia | 932K | 20.2% |
| Florida | 1.9M | 18.3% |
| South Carolina | 427K | 16.7% |
Zipcodes highest index, 2,000+ households
| ZIP | Households | Index |
|---|---|---|
| 85138Maricopa, AZ | 23,719 | 137.3 |
| 34762The Villages, FL | 7,250 | 137.2 |
| 79382Wolfforth, TX | 4,791 | 135.3 |
| 37042Clarksville, TN | 36,818 | 135.2 |
| 80134Parker, CO | 33,408 | 134.9 |
States with at least 50,000 scored households. The largest high-fit audiences by count are in Texas (2.1M), Florida (1.9M), Georgia (932K), California (819K), Colorado (630K).
Metro areas highest high-fit share, 100,000+ households
| Metro area | High-fit | Share | vs U.S. |
|---|---|---|---|
| Clarksville, TN-KY | 63K | 44.5% | 4.43× |
| Greeley, CO | 66K | 41.5% | 4.13× |
| Lubbock, TX | 61K | 40.0% | 3.99× |
| Colorado Springs, CO | 111K | 34.3% | 3.42× |
| Killeen-Temple, TX | 68K | 33.2% | 3.31× |
| Palm Bay-Melbourne-Titusville, FL | 103K | 33.2% | 3.31× |
District of Columbia is a single urban core, so it is shown among states but is better compared with metro areas. The Washington metro area as a whole: 11.0% high-fit across 2.5M households.
Common questions
Where do banking switch potential households concentrate?
By share of households, in the Clarksville, Greeley and Lubbock metro areas. In Clarksville, 44.5% of households are high-fit, 4.43x the national rate. By count, the largest high-fit audiences are in Texas (2.1M), Florida (1.9M), Georgia (932K).
What does high-fit mean for banking switch potential?
A score of 64 or higher, the top 10% of U.S. households for this outcome. Qualified (score 57 or higher) is the top 25%. These are percentile cuts across all scored households, not a count of households that need the product.
Is the banking switch potential score Fair Lending qualified?
It uses no race, sex, age, marital status, familial status or public-assistance inputs, no fields that stand in for them, and no credit bureau data. It is not a credit decision. Zip-level targeting should still get a disparate-impact review before media or offer allocation.
Evidence base data behind banking switch potential
- Semcasting dataSemcasting household file
- Federal statisticsCensus ACS (mobility, internet)
- Regulatory filingFFIEC HMDA 2024
- Regulatory filingFDIC Summary of Deposits
- Regulatory filingCFPB Consumer Complaint Database
- Federal statisticsBLS QCEW
How the score is built
Each household's banking switch potential score (0 to 100) combines its own household fields (50%: home-purchase recency 60%, digital activity 40%) with market signals for its zipcode (50%) from Census ACS (mobility, internet), FFIEC HMDA 2024, FDIC Summary of Deposits, CFPB Consumer Complaint Database, BLS QCEW. Each zip signal is converted to a national percentile. Weights follow Semcasting's acquisition driver model and are refined as campaign and account-opening results come in. State and zip indexes compare an area's average score with the U.S. average (100).
How it is confirmed
Every scored file carries a fixed 10% holdout. As account-opening and campaign outcome data comes in, each score is fitted to actual results and lift is reported by score decile against the holdout.
What it is not
The score is not a credit decision and uses no credit bureau data. It excludes race, sex, age, marital status, familial status and receipt of public assistance, and any field that stands in for them. Zip-level targeting should still get a disparate-impact review before media or offer allocation.