UPST earnings

[LifeOfDreamer]

it all doesnt matter!! The business model has completely changed now… from no loans on the books with no risk to subprime loan with all the risk… aka subprime bank! Reset of valuation and risk management necessary!

---

My understanding is that the R&D loan process is not new.
Another poster mentioned we likely never saw the personal loan version of this, because it likely occurred pre-IPO.

The R&D loan process makes sense. They need to “prove” their models before banks/CU’s will move forward with their products.

I think the drop in price is simply due to the dropped guidance. I think the drop in price is overdone, but I understand there being a drop. Mgmt changing guidance after 1Q, when there isn’t a ton we are seeing today that wasn’t already obvious 3 months ago (Fed poised to raise rates, Russia/Ukraine, end of stimmy checks, etc…) doesn’t fill me with confidence in UPST mgmt’s ability to forecast accurately, let alone in providing consistent beat-n-raises all year long.

You are left with company actually more akin to a value play than a growth play, which isn’t what this board is really about. Granted, if they don’t reduce guidance further in coming Q’s, their 2H of 2022 may be flat sequentially, but it will still be 30%ish growth y/y.

What you have now is a long-game of: what will they be worth IF auto does start executing by EOY or early 2023 as planned, creating a new growth curve. It may very well make the current price in the $30s be an easy double. On the other hand, could they fall further? I don’t know. At a $3b mkt cap, they are still forecasting $1.25b in rev, for a sub-3 P/S.

But if you are working at UPST, and viewed stock comp as part of your package, that isn’t too exciting at moment. So will the company be diluting stock like crazy to retain and attract new talent? Not sure.

Overall, a crappy picture. But I am not worried that much about their loans, as they aren’t trying to be a bank - they are trying to establish data on new products to sell to banks.

I don’t see the point of selling the shares I have at moment, as it is more likely to bounce off a floor here. But not sure I will wait around for the next ER either.

Dreamer

[brittlerock]

Chang - I don’t want to defend the indefensible. Yes, Upstart is a complex story, even more so than I thought. But in the end, IMO, the thing that matters most is, have they actually have developed a disruptive technology? And do they have a significant moat to protect it?

I am reasonably confident that the answer to both questions is “yes.” And if that’s the correct answer, then this current erosion in stock price is likely to be transitory. And let me point out, that our SaaS favorites have suffered a similar fate.

You pointed out that the models appear to be discriminative and that that’s a good thing. But you qualified you assertion by asserting, The only problem with this chart is that I can’t see the sample size for each of the vintage (they could be cherry-picking the FICO segments) and it also doesn’t show year to year trend on whether the model performance has degraded or not recently.

To be blunt, you’re suggesting that the management may well be disingenuous, maybe even somewhat slimy. Of course, we aren’t mind readers and anything is possible. But I would suggest that at no time have been given any evidence that might cause us to question the integrity of Upstart’s management. I concede, I might have missed some things, but I don’t think so. I find it hard to believe that Girouard and the rest of the Upstart C-suite would conceal important information in an attempt to make things “look good.” If that were their game, I would think that this would not be the only item in the quarterly print that may have been “cherry-picked.” In fact, this is a pretty subtle thing which most folks (myself included) would not have even noticed.

So for me, the question is what to do in light of the apparent fact that Upstart will not have another truly high-growth print until a number of macro influences over which they have no ability to alter improve.

In light of the immediate 50% reaction, I intend to do nothing. I believe that it’s not unwarranted, but at the same time, over done, probably driven by a lot of bot sell orders. I’ll be watching for a bounce (am I overly optimistic? Maybe . . .). But, assuming there is a bounce, then what? I’m not sure. As noted, there’s not a holding in my portfolio that has performed will in the last several months. If I sell Upstart, what do I do with the money? Datadog? Already my largest holding, I’m not very comfortable with making it bigger. Start a new position in Mongo? That’s worth considering. I don’t have an immediate answer.

[chang88]

Great point brittlerock, I was not suggesting management being dishonest, but rather a common issue with visual presentations.

The reason I pointed out sample size is because it has direct impact UPST’s bottom line. It is not an easy thing to convey on the chart, but it is easy to have a “chart made of squares” - as we see on page 12, that creates the illusion that each square has the same sample size. But I am pretty sure if you compare the size of A+/>700 FICO group and the size of A+/<600 FICO group, they will not have similar size. The ideal situation for Upstart would be that they found a lot of “A+, low FICO” for their personal loan segment. Sample sizes also matter when you are training to improve models. To dip into low FICO segments to fish out more A+ people in the environment, it seems that they need to take on higher amount of loans on their balance sheets due to rising interest rates. And the 36% cap is limiting their ability to expand the sample size on low FICO segment.

One interesting thought exercise for me was “why the 700+ segment has a 3.4% default rate?” It seems too high. I came to the conclusion from asking myself “why do the 700+ segments need personal loan?” and I think the 700+ here are only referring to people who needs money and have to come to Upstart. This led me to believe that this chat also has a population level bias: we only have data from people who came to Upstart for a loan.

There is much to say and debate here, but the bottom line is I also believe that their product -accurate assessment of risk- works. I don’t think their technology is the problem here. In fact, from what I can find they work great - it’s simply the market condition for this marketplace - matching the seller (banks) and customers (loan applicants) has changed. Like I said, I am putting Upstart high on my radar and plan to reinvest as soon as the next sign of growth show up. The thing is any other competitor needs to go through the same exercise as Upstart if they want to build a resilient model. It’s simply that Upstart is first to the game. And I believe this gave them a tremendous advantage in the long run, much like how Tesla collected tons of real data from reckless users compared to Waymo, whose data came from controlled environments.

On where to put the money: I think that answer is different for everyone. I’m comfortable reinvesting into other companies I currently hold to maximize my time in the market.

[TucsonBones]

Tell me I’m wrong about something

UPST
Today’s Change
$-44.13
$33.00
Previous Close $77.13

You see something worth holding here?

[dividends20]

I bought UPST today. My time horizon is 10 years.

Upstart’s competition, FICO is entrenched, embedded in financials of the society and systems.
It is is simple and effective. Is it ?

The chart that I thought was very interesting is on Page 12.

Upstart has now a competing grade - Upstart Risk Grade (A+, B, C, D, E-) and are now showing head to head against FICO

Management commentary:
Annualized Default Rates as of Mar’22 (Prior 4 Yr Vintages)
Upstart AI continues to separate risk significantly more accurately than traditional underwriting.
https://ir.upstart.com/static-files/abb1aa09-a5a1-48f2-8880-…

But wait, what do their customers say ? That is far more important than what management, analysts or I say.
11 upstart lending partners have dropped all FICO minimum requirements. That number was 0 last year.
They are adding about 1 lending partner a week.
Management: No partners have seen any loan underperformance in light of macro

But wait, there is more.

• Auto lending doing better than expected
• Micro lending in beta and progressing well
• small business lending in initial phases

Watch item: Bridge loans and balance sheet risk. This spooked the markets more than the FY guidance I believe

Q: Sanjay, you mentioned you’ve upped the balance sheet risk here a little bit. How far are you willing to go in terms of supporting new loans and putting warehouse liabilities on the balance sheet?

A: So – and then your second question is how we plan to use our balance sheet. And as I said, I mean, historically, our balance sheet has been almost exclusively for the purpose of R&D. We have used our balance sheet in the last quarter to do what I call sort of a market-clearing mechanism. And by that, what I mean is when interest rates in the economy change quite quickly, I think it would be fair to say that our platform, its ability to react to the new market-clearing price, it’s probably not as nimble as we would like. It’s somewhat manual. It requires a bunch of conversations and phone calls.

And so when interest rates smooth and investors are – so each deciding what their new return hurdles are, there can be a gap or a delay in responding to funding. And that’s a situation where we’ve chosen to sort of step in with our balance sheet and almost sort of bridge to the new market-clearing price. And is that is happening often and abruptly. We’ve been sort of playing that role with our balance sheet.

I don’t view that to be a long-term or necessarily sizable activity for us. I think that developing the mechanisms to respond more nimbly to new price discovery as rates change is something that’s on our road map, and it’s something that we want to start to invest in so that it can happen in a much more automated way. At the end of the day, we view our platform as being a platform that responds to risk and rates in the environment. And so the faster we can do that, the faster we’ll be able to deliver the new returns in any given scenario to the investors and not have to bridge it with their own balance sheet. So I kind of – I would view it through that lens.

[PickTrader]

I was very upbeat about upstart until this earnings report. I do think AI can and will improve lending in a big way and upstart is leading the charge. But, I have large, unanswered questions and have sold out of my position entirely.

1. Banks and auto rooftops are continuing to come online at an impressive rate. It was one of the few impressive metrics from the report. But revenue completely stalled. Why? This is from the Jan-Mar time frame. Interest rates were moving up, but they weren’t going crazy.

2. I’m very worried that the problem will get substantially worse in the future given both (a) increasing interest rates, and (b) the possibility of a recession.

3. Taking non-R&D loans onto the balance sheet seems like another serious issue and it was not well addressed despite many questions on the topic. Being the loans no one would buy, they are obviously the worst loans. But, how much risk is this? How much additional risk can we expect in the future? Are these from the banks, car dealers, or both? I would have preferred management to step up and own the error. Tell us how it happened, how big its scope is, and how/when it will be resolved.

Long term, I think they are in a good place. But near term, I’m not so sure. And, in the back of my mind I worry and wonder about what FICO is doing about this. As magical as AI is, it’s not magic. FICO’s data is not as rich as Upstarts, but they have a lot more of it. And who knows what/how they may doing to get more data.

These are the reasons I exited.

[SaulR80683]

I sold out of my newish Upstart position (third or fourth one, I lost count), not because I thought it might not bounce (it probably will), and not because I thought it won’t be doing well “10 years from now” (I have no real idea, but I suspect that they will find a way to pull out of this nosedive), but because this quarter’s results, and comments by management, both were a surprise huge disappointment. It was way down when I sold but I could buy other companies that were also way down with the money, and they were great companies that I knew were reliable and would do well next quarter, and the quarter after, and probably the quarter after that too. So I took my loss and moved on.

Saul

[dividends20]

But revenue completely stalled. Why?

Cars should slow down when visibility is low.
Insurance companies should insure fewer people when risk is higher.
Lending should tighten when there is economic uncertainty.

In this case revenues going down is desirable assuming the charge rates and quality of borrowers is up.

[FallingWallenda]

Keeping loans on the books knocked me for a loop. That was not why I was invested in Upstart. I was in it for the AI. So many people were saying Upstart is a lender and I thought they were just ill-informed. Duh.

It seemed simple: if Upstart’s AI could increase the pool of qualified borrowers, and reduce defaults - even slightly - financial institutions would have every incentive to use UPST over FICO. Even in a recession, if Upstart AI could prove to be better than FICO, then the lack of overall transaction volume in the industry would be offset for Upstart by the market share they were gobbling up. Now?

Now I am carrying a monstrous 81% loss along with the run-of-the-mill 50% losses on all our other holdings. It seems like a fool’s errand to hold ANYTHING through earnings. Great earnings will get you a yawn, a slight miss will turn you into Charles Manson’s prom date.

[JabbokRiver42]

Brian Feroldi and Brad Freeman did a video dissecting the UPST ER. Brad has done enormous deep dives into Upstart and I thought this was really helpful in understanding where the issues were (and were not) in the report:

https://www.youtube.com/watch?v=d2W6BtaURMM.

The main cause of concern (almost every analyst on the call went back to it) was the steep and abrupt rise of loans held on their balance sheet. About 2/3 of that rise is part of their business model for new loan products. They are holding the newer auto loan segment so that they can generate results to gain the trust of new partners. That’s what they did with personal loans (as someone else has already pointed out, that was before they were public), and I expect they’ll do that with every other new loan category they roll out.

The remaining 1/3 of the increase was the unpleasant surprise and was due to the very fast and very steep rise in yields on the 2-year Treasury. They bookmarked that on the call as the rate to watch most for their business. To keep liquidity with that yield rising so fast, they had to keep more loans on their own books to sell over time. They don’t anticipate that continuing for more quarters; but there are so many macro-events colliding at the moment they were unwilling to rule it out.

Of course the more lending partners they add, the fewer loans they need to sell in the first place, since those partners will originate the loans themselves. That number is going in the right direction, currently adding a new bank/CU partner every week and rising.

I am keeping my shares. I know they will likely be range bound for the rest of the year, but I’m willing to wait for what I believe is a generational company that has not broken my thesis.

JR

[MFChips]

I am keeping my shares. I know they will likely be range bound for the rest of the year, but I’m willing to wait for what I believe is a generational company that has not broken my thesis.

Good post.

In this environment, with so much uncertainty for both loan demand and the ability for UPST to quickly resell them, there is quite a bit of added risk for a stock that we both agree will be range bound for this year, and maybe next.

I think the thesis is intact and this could be a generational company in a few years, but I also think we get 6 months to see if they “prove it”. I sold out a decent position AH last night during the ER call. Dave Girouard lost some points in my book for coming across as shaken and not well equipped to properly communicate this risk to the business, nor properly set expectations all along.

(1) He was selling more stock than usual thru the last quarter.
(2) After they announcement a stock buy back that they didn’t execute whatsoever.
(3) Either he was trying to prop the stock (to sell his shares) or they just didn’t see this balance sheet thing happening and crushing the stock. I’m not sure which is worse.

Upstart would be much better off having a banking partner willing to handle that spread for a hefty fee, which keeps UPST out of the debt/selling business and solely in the Software/AI game. It could be a big margin hit but this is not their core competency. Upstart needs to be a SW company if that is possible, not a blended bank/SW company.

[UptownHoodie]

As JR pointed out, the percentage of loans upstart has to sell is trending down.

From Upstart’s 10Q filed today, total revenue sourced from Cross River Bank has decreased to 46% (down from 56% last quarter! and 60% a year ago):

In the three months ended March 31, 2021 and 2022, CRB originated 56% and 52%, respectively, of the Transaction Volume, Number of Loans. CRB also accounts for a large portion of our revenues. In the three months ended March 31, 2021 and 2022, fees received from CRB accounted for 60% and 46%, respectively, of our total revenue. CRB funds a certain portion of these originated loans by retaining them on its own balance sheet, and sells the remainder of the loans to us, which we in turn sell to institutional investors and to our warehouse trust special purpose entities.

https://ir.upstart.com/node/7866/html

I agree the 25% of non R&D loans on balance sheet is not ideal and is a top item to watch going forward. However, the circumstances appear unique as the market tightened in a matter of days. Yes, upstart should have been better prepared for this scenario but it appears to be an extreme change in environment- fed, war, recession. This is a stress test but upstart navigated the void due to their strong balance sheet which gave them flexibility.

[Texmex]

They don’t anticipate that continuing for more quarters;

It seems so. They also commented they plan to be FCF positive for the rest of the year. This after the huge FCF neg of about -$300M this Q. This suggests they plan to get rid of most of these loans over the next 3Qs. Remains to be seen of course.
Also, note that only 25% of the loan is of concern. 75% of the loan is new product related (mostly auto refi) and management has said numerous times in the past that this is to be expected with new products. No surprises there.

But a lot more interest rate, and macro impact on business than I thought. Also overestimated that they will continue to grow fast by taking market share despite macro - they grew loan originations faster than Lending Club but not by much. Next Q they are projecting the same revenue as LC. Surprised model could not account for the rapid change in the 2-year treasury over the last few months. Most folks are commenting that the 2-year treasury should be more or less stable till the end of 2022 which should work in favor of their model.

This is a much more lumpy business than other companies. Long term the AI model outperformance over FICO and other methods is what matters. Good to see delinquencies returning back to normal.
I plan to hold my shares.

[Texmex]

One more thing. In retrospect I think management was too bullish when they reported in Feb. They were aware that the 2 year had already gone up a lot and deliquencies had been rising and their model had been underperforming. Wish they had been mkre cautious. Hopefully lesson learnt.

[Oracleoo]

One more thing. In retrospect I think management was too bullish when they reported in Feb. They were aware that the 2 year had already gone up a lot and deliquencies had been rising and their model had been underperforming. Wish they had been mkre cautious. Hopefully lesson learnt.

Actually at the point they reported last quarter, based on Trust Pilot reviews as a proxy, they were heading for a Q1 beat as they were likely expecting March to be better than February. 10 year treasury was around 1.7 to 2.0 when they reported, before sharply moving up to 3.0. The drop in Trust pilot review number correlates to sharp rise in treasury. The rise in treasury also which correlates to sharp fall in our stocks.

Should they have anticipated better, yes I think so. But I do think they were taken a bit by surprise by fast 200 basis point rise in interest rates.

[fireblade909]

Should they have anticipated better, yes I think so

By then the stock price had already cratered (I know I kept buying up until the 350s) and I believe management was managing the stock price with hopium to some degree. UPST is my biggest lesson learned in this crash in relation to position sizing and exposure- when the thesis is not working out as well as expected, it’s time to exit / reduce. We read criticisms that Saul et al exit after a brief timeframe of holding … this is why. It’s not that buy n hold is dead, it’s that buy n hold is not appropriate for hyper growth investing. Just in case I sound like I know what I’m doing, please note that I have wiped out my retirement fund that was built from 2019 onwards … the financial and mental health impact is very real

I will hold on to UPST and the rest of my portfolio because price levels are irrational at the moment; and we will have some sector rotation out of consumer staples, energy, etc back in to tech at some point, and it is not as far away as it seems

Long DDOG BILL NET MNDY AMZN

[CMFSircc]

The way Saul handled the Fall in Upstart reminded me of a bible story:

’ … They answered, “Yes, he is dead.”

Then David got up from the floor, washed himself, put lotions on, and changed his clothes. Then he went into the Lord’s house to worship. After that, he went home and asked for something to eat. His servants gave him some food, and he ate.

David’s servants said to him, “Why are you doing this? When the baby was still alive, you fasted and you cried. Now that the baby is dead, you get up and eat food.”

David said, “While the baby was still alive, I fasted, and I cried. I thought, ‘Who knows? Maybe the Lord will feel sorry for me and let the baby live.’ But now that the baby is dead, why should I fast? I can’t bring him back to life. Someday I will go to him, but he cannot come back to me.”

Then David comforted Bathsheba his wife. He slept with her and had sexual relations with her. She became pregnant again and had another son, whom David named Solomon. The Lord loved Solomon.’

---

Reminding me that i don’t have to make money back where i lost it.
Sometimes it is best just to to get up, dust myself off and move on.
Oft times, the new avenue will prove more fruitful than the former.

Best, kevin c