How Rocket Is Using VantageScore to Its Advantage

Mortgage begged for a UAD 3.6 delay. And the GSEs blinked, which means everyone has six additional months to get their 💩 together. I’ve got a full breakdown of what you need to know. 

Also, Rocket is placing a big bet on VantageScore. I’ve got an analysis for paid subscribers detailing how Rocket is tactically using the new credit model to its advantage. Plus, Vishal’s back at Better, a top wholesale lender rebrands & much more!

Well, I guess praying worked 🙏. Back in July, I wrote that the mortgage industry was sleepwalking toward the Nov. 2 UAD 3.6 deadline. (Gift link is here.)

Over the last few weeks, the conversation had basically shifted from, “We’re not ready! Crap, what are we gonna do?!” to, “Surely the GSEs are going to delay this mishegas, right?”

This morning, Fannie Mae & Freddie Mac did exactly that. The GSEs pushed the UAD 3.6 mandate from Nov. 2 to May 19, ‘27, giving lenders, appraisers, AMCs & tech vendors more than six additional months to get it sorted.

And based on what I’ve been hearing lately, they may need every one of them. A truly ridiculous amount of people still aren't ready. Lots of appraisers haven't completed a single 3.6 appraisal. Some lenders are still running legacy LOS or appraisal management systems that aren't fully prepared. And somehow, w/ the original deadline barely a month away, the appraisal software picture had actually gotten worse.

One of the big problems is that lenders are running older LOS & appraisal management systems that still haven't fully adapted to UAD 3.6.

This isn't just an appraiser learning a new form; the entire plumbing has to work. So the lender's LOS has to handle 3.6, the appraisal ordering platform has to send & receive it. Then the AMC has to support it & the appraiser also needs compliant form software. Finally, the finished file has to make its way through UCDP.

And then there’s ACI.

ACI, owned by First American, recently told users that it is discontinuing the current version of its Sky Workbench platform on Oct. 9 b/c, in the company's words, the experience is "not meeting expectations." A significantly improved Sky Workbench 2.0 is targeted for spring ‘27.

Under the old timeline, that was pretty wild. ACI's Sky Workbench was verified by Fannie & Freddie for UAD 3.6 just a few months ago. Then the company announced it was shelving the current version three weeks before UAD 3.6 became mandatory, w/ the replacement not expected until months after the mandate.

Now? Spring ‘27 suddenly fits the calendar a whole lot better.

Meanwhile, Cotality’s TOTAL, the dominant appraisal software platform, has also pushed back portions of its UAD 3.6 development roadmap. Appraisers have complained for months about missing functionality, bugs & workflows that still aren't where they need to be.

Giving the software ecosystem another six-plus months to mature should help. That’s disappointing to firms like Regorra, which says it’s been ready to rock ‘n roll on UAD 3.6. 

Then of course there is the human part. A huge share of appraisers still haven't completed a single UAD 3.6 appraisal. That's important because 3.6 isn't a slightly redesigned 1004. It changes how appraisers inspect properties, collect data, navigate the report & deliver the finished product.

You can watch all the webinars you want, but like, at some point, you actually have to do one? The first 3.6 appraisal completed by one staff appraiser I wrote about this summer took roughly 3x as long as a traditional report. After repetition, he got that down to about 25% longer.

Under the Nov. 2 timeline, a huge chunk of the profession was potentially going to be figuring this stuff out on live loans at roughly the same time. Now they have another six months. The obvious question is whether they actually use them.

There was one thing working in the industry's favor even before today's announcement: business kinda sucks right now 🥹.

In the original July story, I wrote that the nightmare scenario was mortgage rates falling, application volume surging & thousands of lenders, AMCs & appraisers all trying to learn UAD 3.6 at the exact same moment. Instead, rates are closer to 8% than 7%.

Low volume should make this transition easier. Fewer appraisal orders mean appraisers have more time to learn the new workflow, vendors have more capacity to fix problems & lenders can test their systems w/o doing it in the middle of a mini refi boom.

And now they have until May 19, which is good.

The uhh slightly concerning part is that the industry already had a very long runway. Broad production opened in January. And yet, as we approached October, plenty of lenders, vendors & appraisers were still nowhere near ready.

The GSEs have now handed everybody yet another runway. Probably best not to spend the next six months waiting for May, yeah?

On Sept. 16, I asked Rocket’s Bill Banfield where this whole VantageScore thing was headed. He didn’t exactly spoil Monday’s announcement. But he hinted at it.

“With a little over 50% lining up & being similar [to FICO], there shouldn't be anybody hitting a panic button,” Banfield told me. “But I do believe at some point lenders are going to need to decide which path they're going to go down.”

Twelve days later, Rocket picked one. Starting in Q4, Rocket will make VantageScore 4.0 the default on eligible Fannie, Freddie & VA loans, plus anything else that qualifies. Rocket says it’s the first lender to make VantageScore its preferred model.

Except Rocket was already pretty far down this road. My friends at Milliman pulled the channel-level securitization data for me earlier this month. In June, VantageScore represented just 1.4% of Rocket’s retail issuance by UPB. In July, it jumped to 24.1%. August: 44.1%.

Broker went from basically nothing in June to 24.1% in August. Correspondent? Just 5.4%.

So yeah, this is v much a Rocket retail strategy.

Rocket says it has pulled 1.4M dual-score reports this year & that borrowers who saved money using VantageScore saved an average of $1,600 at closing.

IMO, that stat appears to contains a bit of PR magic → The $1,600 figure is the average among borrowers who saved. It doesn’t tell us how many did.

Anyway, Banfield told me VantageScore runs about 17 points higher than FICO on average across Rocket’s files. But a little more than half the time, the scores produce no meaningful difference in pricing or eligibility.