Prequalification: When Yes Turns Into No | HL Hunt
Prequalification: When Yes Turns Into No
Prequalification costs the applicant nothing to ask and the lender little to give, so it drifts toward yes. With a loose threshold, 508 of 2,310 applicants are told they qualify and are then declined, each with a hard inquiry to show for it. Tightening the threshold removes 231 of those reversals at a cost of 99 approvals, 2.3 to 1. The marginal applicants were approved only 30% of the time.
What you'll learn
A prequalification is a promise the lender may not keep
Prequalification tells an applicant, on a soft credit check and a few stated details, that they are likely to be approved. It does not affect their credit file, so it is cheap to ask and easy to market. The final decision comes later, after a full application, a hard inquiry and verification.
The gap between the two is where the trouble sits. An applicant who was told yes and then told no has paid in a hard inquiry, in time, and in trust, and the lender has paid in acquisition cost for a loan it did not make.
How loose prequalification plays out
Take 10,000 prequalification checks. A loose threshold prequalifies 4,200. Of those, 55% go on to apply.
| Threshold | Prequalified | Applied | Approved | Told yes, then no |
|---|---|---|---|---|
| Loose | 4,200 | 2,310 | 1,802 | 508 (22%) |
| Tight | 3,600 | 1,980 | 1,703 | 277 (14%) |
The difference is the 600 marginal applicants the loose threshold includes. They are approved at the final stage only 30% of the time, against 86% for the rest.
That ratio is the decision. The loose threshold funds 99 more loans. It also tells 231 more people that they qualified when they did not, and each of those carries a hard inquiry, an adverse action notice, and a reasonable belief that they were misled. Per our adverse action analysis, a decline after a prequalification also has to explain itself, and the honest explanation is often that the first answer was optimistic.
Why the first answer and the second differ
| Source of the gap | Why it arises |
|---|---|
| Stated income | Per our income estimation analysis, an applicant near a DTI ceiling is decided mostly by estimation noise |
| Soft-pull data is thinner | Some attributes used in the final model are absent from the prequalification view |
| Verification | Per our verification analysis, identity, income and employment checks remove some applicants |
| Time | Per our stacking analysis, new credit taken between prequalification and application changes the picture |
| Policy rules applied only at the end | Per our rules analysis, rules that make most declines are sometimes not run at prequalification |
The last row is the one entirely within the lender's control. A rule that will decline an applicant on information already known at prequalification should be applied at prequalification. Running it only at the final stage produces reversals that served no purpose, since the outcome was knowable when the first answer was given.
A prequalification should predict the final decision
When the same model and rules run at both stages, the told-yes-then-no group shrinks to the cases that verification genuinely changed. HL Hunt AI Underwriting applies consistent decisioning at prequalification and approval, and reports reversal rates by cause.
Why prequalification drifts loose
If tightening is so clearly favourable on the 2.3 ratio, why do thresholds drift loose? Because the costs and benefits land in different places.
- Marketing is measured on prequalified applicants and applications started, which loosening increases.
- Marketplaces and partners rank lenders on prequalification rates. Per our intermediary analysis, the lender that says yes most often gets the most traffic.
- The cost of a reversal falls on the applicant, whose hard inquiry and lost time appear in nobody's budget.
- The cost to the lender is spread out, in acquisition spend on unfunded applicants and in complaints that arrive later.
The commercial case for a looser threshold
Loose prequalification does fund more loans: 1,802 against 1,703 here. If the 99 additional approvals are profitable, and the acquisition cost per prequalified applicant is low, the loose threshold can be the more profitable choice. Some applicants who are marginal on soft data turn out well on full data, and a threshold that excludes them excludes those good outcomes too. The honest form of this argument states the reversal rate it accepts, discloses clearly that prequalification is not approval, and measures complaints. Per our boundary analysis, the applicants near any threshold are where both the gains and the harms concentrate, so the choice is a real trade rather than an obvious one.
Designing prequalification well
- Run the same model and the same rules at prequalification as at approval, on whatever data is available.
- Set the threshold from the reversal rate you are prepared to accept, not from the prequalification rate marketing wants.
- Calibrate on outcomes. Track, by prequalification score band, the share who apply and the share approved.
- Treat the marginal band differently. Tell those applicants they may qualify subject to verification, or offer a smaller amount. Per our counteroffer analysis, a firm smaller offer can beat an optimistic larger one.
- Collect verifiable information early where possible, so the final decision has fewer surprises.
- Make offers firm where the data allows. Per our stipulations analysis, every condition added after prequalification loses applicants, and the good ones leave first.
- Report reversals by cause. Income, thin data, verification, new credit, or rules applied late each have a different fix.
Item four is often the best answer to the trade-off. The marginal applicants are not a homogeneous group to accept or reject; they are applicants for whom the lender lacks confidence. Telling them so, or offering what the lender is confident in, preserves the applicants who will turn out well without making a promise the lender expects to break 70% of the time.
Item three is the measurement most programs lack. Per our point-in-time analysis, calibration has to compare the prequalification score as it was given with the final decision as it was made, and per our small numbers analysis, the bands near the threshold need enough volume to estimate approval rates with any confidence.
Tell marginal applicants the truth before the hard inquiry
Applicants in the marginal band were approved 30% of the time. HL Hunt AI Underwriting identifies that band at prequalification, so it can receive a conditional message or a smaller firm offer instead of a promise that will usually be withdrawn.
Frequently asked questions
Prequalification uses less information. Income noise, thinner data, verification, new credit and rules applied only at the end all open the gap.
Here 22% of applicants with a loose threshold and 14% with a tight one. The marginal band was approved only 30% of the time.
It funds more loans, but here it cost 2.3 reversals per extra approval. If chosen, state the accepted reversal rate and be clear it is not approval.
Consistent decisioning at both stages, thresholds set from acceptable reversal rates, and a conditional message or smaller firm offer for the marginal band.
Key takeaways
- A loose prequalification threshold told 508 of 2,310 applicants yes and then no, 22%.
- A tight threshold reversed 277 of 1,980, 14%, while funding 1,703 loans against 1,802.
- Tightening avoided 231 reversals for 99 lost approvals, 2.3 to 1.
- The 600 marginal applicants added by the loose threshold were approved only 30% of the time.
- Rules that will decline on information known at prequalification should be applied there.
- Thresholds drift loose because marketing and partners reward prequalification rates while applicants bear the reversals.
- A loose threshold can be more profitable, but should state the reversal rate it accepts.
- Give the marginal band a conditional message or a smaller firm offer instead of an optimistic promise.
This guide is educational and does not constitute legal, compliance, or financial advice. All figures are stylized illustrations using assumed prequalification volumes, thresholds, application rates, and approval rates, and do not describe any particular lender or program. Requirements for prescreened and firm offers of credit, adverse action notices, permissible purpose for credit inquiries, and marketing disclosures are governed by federal and state law that varies by product and changes over time; consult qualified counsel before designing a prequalification program.