My Approval Rate Is Already High — So How Can a New Model Help Me?
- Leland Burns & Jim McGuire
- 1 day ago
- 3 min read
A common question we hear from lenders goes something like this:
“My approval rate is already really high. I’m already letting most applicants through — so what’s the point of building a better model?”
It’s a fair question. If your approval rate is 80% or even 90%, the gain from a better rank-ordering model might seem marginal at first glance. But in our work with dozens of lenders across product types, we’ve seen this situation again and again — and we’ve learned that better modeling still makes a meaningful difference, even when headline approval rates are high.
Here’s why.
First, remember what a model is doing
At its core, a credit model is simply rank-ordering risk. A better model gives you more accurate separation — better identification of the safest borrowers and better detection of high-risk ones.
You can turn this accuracy into value in a few different ways:
Approve more people at the same loss rate
Lower losses while keeping the approval rate steady
Or — and this is where it gets interesting — optimize loan terms and operations within the approved population
This last use case is where most of the upside comes from when approval rates are already high.
Not all approvals are equal
Let’s take a step back. When we hear “85% approval rate,” what does that really mean?
In practice, a large chunk of that 85% may not be “straight approvals.” They might instead be conditional approvals — offers that come with high rates, counteroffers, or manual stipulations. These approvals often carry:
Loan terms that few borrowers accept
Friction from extra steps or manual review
Low conversion rates
That’s especially true in indirect lending environments like auto finance or home improvement, where the lender isn't controlling the direct borrower interaction and must compete with other financing options being offered simultaneously.
A better model won’t just increase the number of people you approve — it will help you:
Approve the right people with the right terms
Push more borrowers into your most competitive pricing tiers
Reduce manual steps and conditional offers
Improve conversion and operational efficiency
In other words, within your high approval rate, there's room to improve performance.
Real-world examples: auto and solar lenders
We’ve seen this pattern repeatedly.
In indirect auto, clients often approve 75–85% of applicants. But many of those “approvals” are actually low-tier counteroffers that few borrowers take up. A better model allowed one client to confidently move a larger share of borrowers into their most competitive pricing tier — doubling conversion in that group, without increasing credit risk.
In emerging market auto lending, one lender maintained an already-high approval rate but saw dramatic improvement in conversion after implementing a machine learning model. They could offer their best terms to a more precise subset of the population, boosting funded loan volume and borrower satisfaction.
In home improvement lending, we worked with a client serving prime borrowers shopping for solar loans. These customers were highly rate-sensitive and had multiple competing offers. By improving the model, we helped the lender identify the best-risk borrowers more confidently — and offer them more competitive terms. Again, conversion surged.
Even a 90% approval rate means 10% denial — and that matters
Don’t forget: even if your approval rate is 90%, you’re still denying 10% of applicants.
That bottom decile matters. If you're going to reject people, you want to be sure you're rejecting the right ones. And modern machine learning models are especially good at identifying the worst risks in a population — improving who you reject, even if the number you reject doesn’t change.
That means lower losses from better bad-loan avoidance.
Final thought
If you’re sitting on an 85% approval rate, it’s natural to ask whether a better model can really help you.
But from what we’ve seen, the answer is almost always yes.
A better model isn’t just about approving more people. It’s about converting more of the right people, under better terms, with less friction — while also rejecting the worst risks more confidently.
High approval rates don’t mean you’ve maxed out your modeling ROI. In fact, they often mean you’re sitting on untapped gains.