Upstart Case Study

Upstart published a very interesting case study. Now, I recommend taking any statistics directly from a company with a grain of salt. But, this could be telling about Upstart’s ability to create value for customers and to expand the market to new borrowers. We are still in the early innings of the Upstart story.

The highlights:

3 Years of Partnership in Review

-Scaled loan volume target from $500k/month in April 2019 to $12M/month as of March 2022?
-Increased max loan size from $20-30k to $30-$50k
- >1,000 new customer relationships as of February 2022?
-Grew from 24.5% lending distribution to LMI communities to 38%?
-Expansion into Auto Refi in Q2 2022

While I like the way this data looks, it is a single case. We do not know what other variables might have affected these numbers and we do not know the details for every other bank. I am eager to find out, though.

*LMI = low to moderate income

https://info.upstart.com/ffbkc-case-study

45 Likes

Good share, StocksAndStouts!

This case study illustrates for me the flawed logic of arguments that say banks or fintech will just come up with their own AI and won’t need Upstart to get to the same level.

This bank just 24x their personal loan volume in 3 years, to $12M, and all they had to pay for that was about 6% of that loan volume, or about $720k… another telling number is the LMI communities increase of almost 60% in loaning to people that would not have been even in the conversation with FICO scores or traditional risk models for personal lending. i don’t think they are getting to the same spot by hiring a few data scientists and programmers and asking them to feed some stats into a model for lending :slight_smile:

Upstart spent about 134 million in R&D last year, and 40 million the year before - I think banks are getting a deal using Upstart!

JDR

19 Likes

3 Years of Partnership in Review

-Scaled loan volume target from $500k/month in April 2019 to $12M/month as of March 2022?
-Increased max loan size from $20-30k to $30-$50k
- >1,000 new customer relationships as of February 2022?
-Grew from 24.5% lending distribution to LMI communities to 38%?
-Expansion into Auto Refi in Q2 2022

This is the sort of broad brush info that has VALUE.

Our companies need to be doing things very well, but they don’t need to be perfect when examined with a flawed* microscope. The best don’t bat a thousand. Not even 0.900

*Flawed: Detailed analysis on Upstart invariably lacks full information as to what is going on AND it invariably fails to know the intent of the lending institutions and of Upstart (including trade-offs on loan acquisitions vs rates vs losses. IMO, we’ll just have to muddle along with triple digit performances without complete understanding.

Not a frightening situation, IMO. If you’re sailing toward a fair harbor, it’s OK if the waves don’t remain precisely lined up. Variability is reality.

Rob
Rule Breaker Home Fool
He is no fool who gives what he cannot keep to gain what he cannot lose.

29 Likes

What would be interesting to me to contextualize this is what is a typical bank margin on a loan. Does contracting with Upstart increase that margin on net because of better repayment rates, or does the cost of the contrast decrease that margin and require the bank to make up that income on volume? Without that context, the simple fact of increased loan volume doesn’t seem super-relevant.

2 Likes

The earlier thread on KBRA reports highlighted to me the fact that FICO scores are entrenched, tried and trusted.
They represent a good combination of simplicity, effectiveness and risk in one simple number.
FICO scores are a part of the financial system and the society.

The Upstart investment thesis is based on ones conviction if Machine Learning (ML) models will outperform based on a wider variety of metrics.
I believe the ML models will do better but it will be a bumpy ride. Disrupting such an entrenched system is not simple. It is going to be two steps forward and one step back.
The other factors (ease of use, speed of decision, consumer satisfaction) are all favorable to Upstart and I don’t think there is any debate there.

8 Likes

Thanks for posting this case study. Interesting stuff that well highlights Upstart’s strategy and value.

From the story: https://info.upstart.com/ffbkc-case-study

"
FFBKC was able to go live with Upstart in just three months and was pleased with Upstart’s transparent and consultative approach…Cooper [Chief Innovation Officer at FFBKC]…emphasized that Upstart’s dedicated Customer Success and Account Management teams were a huge factor in deciding to scale its personal loan targets.

During the pandemic, the Upstart team met weekly with FFBKC to share trends Upstart was observing from other lenders and ensure that the bank was achieving its loan targets.

This period also proved Upstart’s capability to quickly ramp up FFBKC’s loan volume target. In just two and a half weeks, FFBKC was up and running with a higher loan volume…So, what’s next on the horizon for FFBKC? Auto refinance.

“Auto lending was just the logical next step in growing our relationship. We want to grow a relationship that is very cohesive, transparent, and valuable to the bank,” Cooper said. “When you find a winner, you go all-in.”[my emphasis]
"
Cooper is also quoted as being open to adopting other Upstart products in the future since the partnership has gone so well so far. So, when Upstart rolls out small business loans in the next year or two, they’ll find a receptive audience among banks like this one who’re already customers.

It’s interesting how unpopular UPST stock has been recently. Profitable + hyper-growth + huge TAM and yet on an up day for NASDAQ upstart is ~2.5% in the red. Upstart is in that awkward phase of transitioning from its stock being valued purely on growth and possibilities to being valued more on its proven earnings. As an investor who is far more concerned with valuation that typical on this board, I am watching the price closely and hope to enter soon if the decline continues.

I’m not much of a believer in technical analysis but can’t help thinking along those lines sometimes. UPST’s lowest close since its price blew up last year has been $89.34, with several forays down to roughly that level. Today it looks likely to close below that, breaking through support at 89-90. The lowest it has traded in recent months was a brief spike down intraday that touched $75.15 on January 24th. PE at current price is ~62, and that would drop to ~51 if it reaches $75 again. I think I’ll be buying if it gets that low.

I know that the last paragraph is going OT to this board, hope it’s not too much. Feel free to delete if it’s too far over the line.

5 Likes

It’s interesting how unpopular UPST stock has been recently. Profitable + hyper-growth + huge TAM and yet on an up day for NASDAQ upstart is ~2.5% in the red. Upstart is in that awkward phase of transitioning from its stock being valued purely on growth and possibilities to being valued more on its proven earnings.

BenSolar,

My take on UPST’s Q1 2022 is purely speculative, but here is what I’m seeing today.

Recall that Trustpilot review count appears to track closely with loan transaction numbers since July 1, 2021. Starting July 1, 2021 a ‘switch was flipped’ where they generated far more automated invitations for Trustpilot reviews, which is how fee revenue numbers have closely correlated since then.

I’ll paste from my prior post (https://discussion.fool.com/upst-q4-8211-great-quarter-35054285…:slight_smile:

“Trustpilot review count appears to correlate quite well with loan transaction numbers:
6080 reviews in Q3 with 362780 loans. That’s 59.7 loans per review.
8336 reviews in Q4 with 495205 loans. That’s 59.4 loans per review
At roughly halfway through Q1 (February 15 2022), there were 4236 reviews.
4236 x 2 x 59.55 = maybe 504507 loans for Q1.

The total Trustpilot review count for Q1 2022 sits at 7505. Multiplied by 59.55 = estimation of 446923 loans transacted. This would be below my earlier estimation written above at 504507.
If we assume (lots of assumptions here) that loan average size stays unchanged at $8275 and take rate at 7.0%, we calculate total fee revenue of $258.9M. In order for UPST to meet the high end of total revenue Q1 2022 guidance of $305M, they will need to produce $46.1M of interest income/securitization/loan sale income. However, in Q4 2021 they only generated 18M of this type of income, so to get there they need to increase that by 356% QoQ, which might not be realistic.
***at end of this post reply, I also compare UPST to competitor Upgrade (who also offers personal loans) trustpilot review counts to help ‘validate’ the QoQ drop in activity

The conclusion is that if (always a big if) the trustpilot counts remain an accurate prediction of loan transaction numbers for Q1 2022, it is highly likely that UPST will miss their revenue guidance and have a sequential drop in total fee revenue by -9.8% QoQ.

Now the major caveat here is that we know UPST management spends their time reading this board too, so it’s very possible that they tweaked the review invitation mechanism just to throw this alternative-data measurement off ever since then.
In any case, it’s possible other investors out there are seeing this same possibility above and that can partially explain share price reaction since Q4 2021 earnings.

Other things to note:

  1. as I wrote before, auto refi numbers were confirmed as a disappointing drop QoQ from Q3 to Q4 based on their 10K filing (https://discussion.fool.com/the-numbers-of-2000-and-4000-from-pr…) This may not bode well for the current quarter or rest of year 2022, especially when you include the fact that interest rates for auto have risen, so auto refi offers may not be as attractive to consumers as a year ago.

  2. Wall Street Journal article yesterday also pointed out the rising yield spread for consumer loan debt and called out UPST and AFRM (https://www.wsj.com/articles/investors-turn-cautious-on-cons…:slight_smile:
    “This year, investors have sold bonds broadly, driving up yields, which rise when prices fall. But consumer-debt yields are rising even faster, a sign that traders believe the relative risk is increasing. Bonds backed by the most-traded category of subprime auto loans have recently yielded 1.45 percentage points more than standard benchmarks, according to data from JPMorgan Chase & Co. , up from a 0.9-percentage-point premium, or spread, at the start of the year. Yields also have climbed for bonds backed by credit-card debt and other types of consumer debt…Shares of Affirm and competitor Upstart have each lost about 75% since November when late payments started to rise, according to FactSet. Short interest as a percentage of shares outstanding has tripled for Upstart to about 15%A spokesman for Upstart declined to comment.

  3. Credit card outstanding balances have been falling since end of 2021, according to MoodyAnalytics/CreditForecast. Consumer borrowing has shifted away from credit cards (see graph here: https://twitter.com/ayeshatariq/status/1513448476796989442?s…). Recall that much of UPST’s personal loan business is reliant on credit card refinancing into a personal loan. It’s possible other UPST investors are seeing this negative datapoint trend and can explain some of the share price fall.

***comparing UPST to Upgrade trustpilot counts:

            UPST    month/month change
Jan count   2919    -7.8%
Feb count   2158    -26.1%
Mar count   2428    +12.5%
            Upgrade    
Jan count   1125    -27.9%
Feb count   1040    -7.6%
Mar count   1115    +7.2%

The appearance here is that the trends in month to month counts is the same general direction for UPST and competitor Upgrade and to me makes it more likely that the review counts are a somewhat accurate reflection of personal loan demand for UPST

27 Likes

perhaps a Q1 revenue miss is why Dave sold shares.

https://www.wallstreetzen.com/stocks/us/nasdaq/upst/ownershi…

Upstart is very bad at pumping there stock. They could have already bought back million’s of share’s to support the stock price, but chose not to. Maybe another piece of information to read into.

Imagine if they un-authorize the sharebuybacks. Negative-pumping on the stock price.

(Long UPST…)

1 Like

@str1der

Your post reveals a lack of knowledge around what execs can and cannot do. An executive like Dave G. has basically “no ability” to sell his shares outside of an SEC registered 105b-1 plan. He could buy outside of a “plan” but selling as long as he is an executive with inside knowledge is basically not possible. The registered plan sets a schedule for shares to be sold, so that the exec has no control over when they get sold, nor any control over how many get sold.

Also, we have no way of knowing whether or not Upstart has been buying shares under the buyback authorization, until they report. There are no SEC documents for them to file “as they go”. So when they report the next quarter, they will reveal their uses of funds during the quarter. That is the only time we’re going to find out about the share repurchases, if they happen.

Hope that helps you maintain a more level head, regarding what is and is not going on, i.e. what to worry about, while the share prices drop. Things are dropping right now because investors are worried about what they don’t know about the future (on many levels).

-Another Rob

27 Likes

Interesting analysis, jonwayne. I’ve thought about this and I suspect there is some noise in a straight line correlation between trustpilot and UPST loan volume. I noticed Upgrade maintained a consistent 36 reviews/day for the entire Q whereas UPST dropped off from 94 reviews/day in Jan to 78 in Feb and Mar. It’s possible this reflects a rapid deterioration in the business starting in mid Feb. I looked at semrush to what UPST site traffic was like over the period and in fact overall web visits were down sequentially month over month during the Q (not so for Upgrade, which were up sequentially m/m over the Q). BUT, site visits for UPST were up in 1Q22 versus 4Q21 by roughly 4m or 17%. To end up with the loan volume you’re projecting, UPST would have to convert visitors into borrowers at a materially worse rate in Q1 versus Q4, roughly a 30% decline in conversions.

Google trends showed a continuous up-trend through Q1 for “Upstart loan.” There appears to be at least some continued interest in Upstart from among would-be borrowers.

Against all this data, you have to assume that the business deterioration over the second half of Q1 was so rapid and so unexpected that the management team couldn’t foresee it on Feb 15 despite having seen financials for the fist half of the Q and having a view into the borrower pipeline over, let’s say, the next 5 days (it takes 1-10 days to get approved for an UPST loan).

31 Likes

Seems like people who are either all in or all out of UPST always come up with something. Smart people new to investing, even those who are outstanding at secondary research and generously forthcoming, are not necessarily good at predicting the future. It’s a common dichotomy of talent and a reality i learned to use to my advantage as an executive.

I thought the KRBA delinquency pop was effectively refuted in the last earnings transcript by the UPST leadership.

Now we hear there is reason to be concerned that online review counts may predict a revenue miss.

Maybe so. These review counts may tell us something important, but not sure it will result in a revenue miss. I doubt it.

UPST CEO’s predicted pace of 100+ bank partners by this summer still looks intact. Wouldn’t all of these loans be of the white label type, and not subject to UPST reviews? Might this factor be a compensating revenue source? Didn’t we agree last year that UPST needed expanding revenue sources to maintain growth for the longer term?

There are likely other changing variables that will work in both directions. Upstart is a young rapidly expanding business with multiple changing verticals. We all expect, indeed hope, that a lower percentage of loan originations will occur from Credit Karma.

I don’t expect a big upside surprise in auto loans this quarter, but i think the odds of a revenue miss are low. We were warned of lumpiness in the revenue model but i’ll be surprised if UPST misses in Q1.

My UPST holding is down to 6%, still substantial, but not the 20%+ i held last year. My first buy was January 6, 2021 at $47 and my returns on the stock have been very satisfactory.

I’m wrong a lot.

Best to all.

57 Likes

It is important to not over think and over analyze. Just read the press release.
It is hard to predict stock price and explain stock price movements.


https://ir.upstart.com/news-releases/news-release-details/up…

Fourth Quarter 2021 Financial Highlights

Revenue. Total revenue was $305 million, an increase of 252% from the fourth quarter of 2020. Total fee revenue was $287 million, an increase of 240% year-over-year.
Transaction Volume and Conversion Rate. Bank partners originated 495,205 loans, totaling $4.1 billion, across our platform in the fourth quarter, up 301% from the same quarter of the prior year. Conversion on rate requests was 24% in the fourth quarter of 2021, up from 17% in the same quarter of the prior year.
Income from Operations. Income from operations was $60.4 million, up from $10.4 million the prior year.
Net Income and EPS. GAAP net income was $58.9 million, up from $1.0 million in the fourth quarter of 2020. Adjusted net income was $87.0 million, up from $5.4 million in the same quarter of the prior year. Accordingly, GAAP diluted earnings per share was $0.61, and diluted adjusted earnings per share was $0.89 based on the weighted-average common shares outstanding during the period.
Contribution Profit. Contribution profit was $149.5 million, up 261% from in the fourth quarter of 2020, with a contribution margin of 52% compared to a 49% contribution margin in the same quarter of the prior year.
Adjusted EBITDA. Adjusted EBITDA was $91.0 million, up from $15.5 million in the same quarter of the prior year. The fourth quarter 2021 adjusted EBITDA margin was 30% of total revenue, up from 18% in the fourth quarter of 2020.
Full Year 2021 Financial Highlights

Revenue. Total revenue was $849 million, an increase of 264% from 2020. Total fee revenue was $801 million, an increase of 251% year-over-year.
Transaction Volume and Conversion Rate. Bank partners originated 1.3 million loans, totaling $11.8 billion, across our platform in 2021, up 338% from the prior year. Conversion on rate requests was 24% 2021, up from 15% in the prior year.
Income from Operations. Income from operations was $141 million, up from $11.8 million the prior year.
Net Income and EPS. GAAP net income was $135 million, up from $6.0 million in 2020. Adjusted net income was $224 million, up from $17.5 million in the prior year. Accordingly, GAAP diluted earnings per share was $1.43, and diluted adjusted earnings per share was $2.37 based on the weighted-average common shares outstanding during the period.
Contribution Profit. Contribution profit was $398 million, up 279% from 2020, with a contribution margin of 50% compared to a 46% contribution margin in the prior year.
Adjusted EBITDA. Adjusted EBITDA was $232 million, up from $31.5 million in the prior year. Full year 2021 adjusted EBITDA margin was 27% of total revenue, up from 13% in 2020.
Financial Outlook

For the first quarter of 2022, Upstart expects:

Revenue of $295 to $305 million
Contribution Margin of approximately 46%
Net Income of $18 to $22 million
Adjusted Net Income of $50 to $52 million
Adjusted EBITDA of $56 to $58 million
Basic Weighted-Average Share Count of approximately 84.3 million shares
Diluted Weighted-Average Share Count of approximately 95.9 million shares

For the full year 2022, Upstart expects:

Revenue of approximately $1.4 billion
Contribution Margin of approximately 45%
Adjusted EBITDA of approximately 17%
Auto Transaction Volume of approximately $1.5 billion

16 Likes

We have now had 12 posts analyzing an Upstart Case Study. That is way, WAY, over-analyzing and nit picking little details instead of the company’s basic numbers, and this kind of focus on little details got us in trouble once before with Upstart. Let’s drop this thread, or I’ll have to start deleting future posts. I don’t want to have to do another two week ban on it. I really don’t!

Saul

30 Likes

I suggest you read the Monday Morning Rules of the Board post above from Saul. (I couldn’t see one from yesterday, but there’s definitely one from 25th April.

Alex

4 Likes