Upstart: a tale of two stories

There has been a narratives going around, which many - on the board and off, judging by recent share price performance - believe. I will try to debunk, or at least paint a different possibility to, this prevailing thinking. Boiled down, the story goes like this:

The bearish story

Upstart’s personal loan product has seen such incredible growth in 2021 that it has rapidly run the course of its S-curve and is now plateauing. This is further supported by the fact that in Q3 they were already at a market share of 15% of their TAM [$3.130bn / ($84bn/4)], and this is likely not far from the max they can achieve. Their rapid qoq revenue growth deceleration gives further credence to the fact that their personal loan product has run its hyper-growth course. They therefore need another product to maintain hypergrowth, which they have duly tee-d up in the form of auto. Unfortunately, though, their auto product will not materially contribute anytime soon. Their Q4 is therefore not likely to be much above their guide of $265 - let’s say $275m - which will give them a total revenue for the year of $818m. Given the above we should be happy for a guide for 2022 at or slightly above $1bn, for a paltry 22% yoy growth, and ecstatic for anything higher. The combination of relatively uninspirng Q4 growth and a poor guide for 2022 will keep the stock depressed.

There are two very important assumptions in this story that I would like to unpick in this post:

  1. They’ve reached about 15% of their personal loan TAM already and that is close to a plateau
  2. Auto will not materially contribute anytime soon

I will first critically evaluate these assumptions and then offer an alternative narrative, which is much more bullish.

By the numbers first, the stuff we know:

**Rev $m	Q1	Q2	Q3	Q4**
2018				24
2019	20	33	49	62.6
2020	64	17	65.4	86.7
2021	121	194	228.4	**285.3e**

	YoY	YoY	YoY	YoY
	Q1	Q2	Q3	Q4
2019				161%
2020	220%	-48%	33%	38%
2021	89%	1041%	249%	**229%e**

The Q4 numbers are based on my attempt to guesstimate Q4 numbers, here:…

Underpinning the revenue are the following number of loans and $ value originated:

#	Q1	Q2	Q3	Q4
2019	30,998	41,211	64,259	78,654
2020	84,214	11,873	80,893	123,396
2021	169,750	286,864	362,780	**453,198e**

QoQ	Q1	Q2	Q3	Q4
2019		33%	56%	22%
2020	7%	-86%	581%	53%
2021	38%	69%	26%**25%e**

$m	Q1	Q2	Q3	Q4
2019	369	486	825	1,044
2020	1,123	164	909	1,249
2021	1,729	2,795	3,130	3,910

I will use the above numbers for Q4 in the analysis that follows.

How big is the US unsecured personal loans TAM?

Upstart stated that their TAM was $92bn in Q1 2021 and then reduced it to $84bn from Q2 onward. It is based on Transunion data. So that’s where we need to dig to understand their TAM and market share better.

First thing to notice is that both the $84bn TAM is calculated by added 4 quarters of Transunion data, namely: “Total unsecured personal loans using data provided by TransUnion for Q2 2020 – Q1 2021 “

Why is this relevant? Because their TAM is made up of 4 quarters: Q2 2020 was the first quarter of the pandemic, during which loan volumes and value literally fell off a cliff and the 3 quarters after that, when the market was similarly depressed. So the $84bn is understated. What about the $92bn which they gave in Q1? “Total unsecured personal loans using data provided by TransUnion for Q4 2019 – Q3 2020.” So including 2 COVID quarters. So the $92bn is also understated.

Key take-out here is that the unsecured personal loan market is quite a bit bigger than $84bn. Using the same Transunion data we can calculate that In 2019 it was $117bn - $33bn higher!

What is Upstart’s actual market share?

Luckily Transunion also publishes quarterly data. Using the reports below from Transunion, we can tabulate both the volume and value of unsecured personal loans from Q1 2019 to Q2 2021

Q4 2020:…
Q1 2021:…
Q2 2021:…
Q3 2021:…
Q4 2020:…

Market volume and value

**Vols m	Q1	Q2	Q3	Q4**
2019	3.8	4.8	5.1	5.2
2020	3.9	2.6	3.5	4.2
2021	3.2	4.4	4.8	5.1

**Val $m	Q1	Q2	Q3	Q4**
2019	25,316	30,202	32,008	29,387
2020	25,861	15,558	20,356	21,738
2021	22,813	31,539	34,319	32,647

All numbers up to Q2 2021 are actual. I forecasted the market conservatively further for Q3 and Q4 of 2021 which allows the calculation of the actual Upstart market shares - both volume and value per quarter:

Market share

**Vol	Q1	Q2	Q3	Q4**
2019	0.8%	0.9%	1.3%	1.7%
2020	2.2%	0.5%	2.3%	3.0%
2021	5.3%	6.5%	7.6%	8.9%

**Val	Q1	Q2	Q3	Q4**
2019	1.5%	1.6%	2.6%	3.6%
2020	4.3%	1.1%	4.5%	5.7%
2021	7.6%	8.9%	9.1%	12.0%

What we see above is that Upstart’s market share has been steadily inching up from around 1% beginning 2019 and improving by a couple of percentage points in 2019 and 2020 before rapidly accelerating in 2021. Something changed to propel their more recent market share gains which is mainly that their product became much better and the market - both banks and consumers - became more receptive. However even with all of that rapid growth, the table above shows that their market share is actually much lower than the 15% bandied about, because the market has grown a lot from the depths of the pandemic, which is what Upstart has chosen to use as its TAM.

So the second key take-out is that Upstart’s market share is not already at 15%; it was only 6.5% of volumes and 8.9% of value in Q2 2021 and, based on my assumptions will be 8.9% of volume and 12% of value in Q4.

So what so far?

Both the fact that the TAM is actually much bigger than the TAM published by Upstart and the fact that Upstart’s actual volume share of that market is not even at 10% yet are important. For some reason, there is an assumption that 15% of the market is some kind of ceiling for them, after which surely hypergrowth would stall. Not only is that assumption flawed - why on earth is 15% a ceiling given the growth seen thus far, driven by their superior product? Why is 10% not a ceiling? Or 20%? Or 30%? I contend that we don’t know, but given the momentum in the business we’re not there yet. But also the 15% itself is incorrect: they are, in fact, a long way from reaching even that market share.

My conclusion is therefore that they have a long runway in personal loans left - they still have a sub-10% market share in a >$100bn TAM.

What about 2022, though?

From what I’ve read about accelerating credit card originations (here: and the loosening credit card standards ( published by the Fed, it is clear to me that 2022 is expected to be a better - much better - personal loan market than 2021.

Transunion also feels that 2022 is shaping up to be a bumper year for unsecured loans, and that a lot of the growth will be driven by Gen Z and Y. Now remind me again, wasn’t that exactly the market that Upstart is disproportionally strong in? (…)

Here’s a key quote from the above:

“Originations for personal loans are expected to continue rising in 2022 – marking seven straight quarterly increases. Originations to both non-prime and prime and above consumers are expected to grow in 2022 and resemble pre-pandemic origination volumes – levels last seen in 2019.”

So the third key take-out is that 2022 is expected to be, and the trends are pointing to, a much better personal loan market than 2021, and that the outsized growth will be driven by the very customers that Upstart is great at targeting. In fact, using Transunion’s expected loan volumes and conservatively assuming lower average loan sizes than this year yields a personal lending TAM for 2022 of $132bn.

And Auto loans?

The CFO told us that auto will be material in 2022; were we not listening? Here is the quote from Q3:

“we believe Auto will be a meaningful contributor to our financials next year, but we don’t have specifics on what exactly that means”

So I’m not going to re-invent the wheel (see what I did there?) I’ll go with jonwayne’s auto revenue estimate for 2022:…

He estimates that they could do $192m in auto loans in 2022. Now, at first glance that could seem small, but that’s a bit lazy thinking imo. We need to factor in that they are coming off a very low base, so need to ramp that up over the 4 quarters.

Let’s say, something like this from Q1 to Q4:

Auto loan revenue: $12m → $25m → $55m → $100m.

If we assume that jonwayne is in the ballbark and that we need to phase in that $192m revenue over the 4 quarters, starting from close to zero, then by Q4 next year, auto could be contributing almost 20% of their revenue.

And their 2022 guide? Surely that is going to disappoint?

Not so fast. Yes, Q1 is seasonally not as strong as Q4 in more normal years (it dropped sequentially from 2018 to 2019 and was flat from 2019 to 2020). So let’s assume that will again be the case - a relatively flat Q1 2022 sequentially. And let’s also say that I’m close to the truth and they report around $285m in Q4 2021.

Then, using a simple x4 Q4 run-rate, they could safely guide to $1,140m for 2022, which would be good for a 37.7% yoy growth rate, leaving ample room to beat and raise. That will certainly be one way that they will look at their guide for 2022.

But another way will be to actually model out what they could expect to hit if all went well, and then guide lower.

So let’s give that a bash. Let’s assume that they can continue growing their personal loan volumes in 2022, but at decelerating pace of growth. Let’s assume qoq growth from Q1 to Q4, starting at my estimate of 453k loans in Q4 2021, as follows:

QoQ loan growth: 0% → 21% → 18% → 15%

So no growth in loan vols in Q1, and let’s also reduce the average loan values from around $8.6k to around $8.3k over the same period.

That gives us origination volumes and total loan values as follows:

Orig	Q1	Q2	Q3	Q4
2019	30,998	41,211	64,259	78,654
2020	84,214	11,873	80,893	123,396
2021	169,750	286,864	362,780	453,198
2022	453,198	549,713	646,228	742,743

$m val	Q1	Q2	Q3	Q4
2019	369	486	825	1,044
2020	1,123	164	909	1,249
2021	1,729	2,795	3,130	3,910
2022	3,898	4,673	5,428	6,165

So, for personal loans that will result in them doing $6.165bn of loans in Q4 2022, out of an expected market of $33.4bn. for an 18.4% market share end of next year.

Let’s also assume that they take a smaller cut per loan and that the average revenue per loan drops from $630 down to $600 over the four quarters.

Putting that in our pipe, adding auto loan revenue and smoking, gives the following revenue and growth estimate for 2022:

Rev	Q1	Q2	Q3	Q4
2022	297.5	365.8	449.2	545.6
%yoy	146%	89%	97%	91%

Or, a total of $1.658bn Revenue for the year 2022 or up 100% vs the $828m for 2021.

Now I’m NOT saying that will happen. Hell, who knows exactly. But the above is a possibility if all goes well, just as much as the dropping off a cliff narrative that I started this post with is a possibility - although with a lower probability imo.=

A guide of $1.1bn would give them 33% yoy and $1.2bn 45% yoy, with room to beat and raise. If they were in stead to guide for $1bn, then, as others, I would be disappointed as this would be only 20.7%, and that would signal they don’t have as much confidence in a scenario like the one I outlined above coming to pass.

Therefore a guide of between $1.1bn and $1.2bn would be what I expect and would like to see as I believe they would then be setting themselves up for a credible path to a beat and raise sequence for the year which could come close to the $1.5bn mark in the end, or 81% yoy growth.

I’m putting my money on a $1.15bn guide for next year, and flat qoq revenue growth guide for Q1 (which would still be good for 136% yoy).

Which brings me to the bullish story:

Upstart has been beaten down unnecessarily and the market is under-appreciating both the auto opportunity and the longevity of their personal lending product. In Q3 of 2021 they grew their loan originations by 26% qoq and the momentum in the business is sure to keep them growing at pace for many quarters to come in personal lending. In addition there are clear signs of improving credit conditions for lenders, in personal loans as well as credit cards, which is the key debt that they tend to replace, and for Gen Z and Y which are their core target market. Both trends bode well for Upstart. In addition, their auto lending product will start to meaningfully contribute in 2022 - adding another engine of hypergrowth to the first and may already contribute up to 20% by the end of the year.

Their Q4 is likely to produce a substantial beat - coming in at around $285m and setting them up for a very strong guide for 2022 despite a flat guide for Q1, which could very well ignite the stock.

Pick your story. I’m picking the second.


(Long UPST).


Another reason the TAM estimates based on Transunion may be low is that by definition of what Upstart does, they permit some people to get personal loans who never would have been able to get approved to get them before. That has to increase the TAM.

5.6% position in Upstart (which is a small fraction of the one I once had, but still significant).


Been noticing headlines lately like;

“Lending to borrowers with low credit scores could increase in 2022, TransUnion forecasts”

“Auto, credit card and personal loans to expand low-credit-score offerings”

“Americans are racking up credit-card debt again — as mortgage forbearance ends and prices continue to rise”

“Americans look to credit as stimulus money dries up”

Seems to be serving up a fastball down the middle of the plate for UpStart.


Upstart can expand into “insurance” risk / premiums.


in post 82129 last week I tried to make the very same point, and take a stab at calculating that TAM:

From UPST website - Less than 48% of Americans have access to prime credit (read: FICO score enabled) - but 80% of Americans have never defaulted on a credit product.

So they are targeting the 4 out of 5 that will pay in the half of Americans that would payback but don’t even have the chance to get approved by traditional lenders.
There are 210 million adults in the US, so 0.8 x 105 = 84 million as a target
A CNBC article on 2020 mentioned that nearly one quarter of Americans (22%) have a personal loan (let’s not count auto or mortgages yet)and that the average being $16k in balance
Personal loan balance underserved possible TAM = 84M x 0.22 x $16000 = 296B
maybe they don’t want to borrow as much, so let’s say that this underserved market half of that average balance.
TAM is then 148 billion dollars

So when UPST tells you that the personal loan TAM is 84B, they are underselling it - :slight_smile:



Does is really makes sense to underselling there own TAM? All fanboys jump in when they see a really big TAM, it looks great in the prospects and so on…

Normally companies making there TAMs much bigger then it really is. And that’s basically the reason why they never reach over 10-15% of the TAM :slight_smile:

Ok a growthing TAM from small base looks great and makes positive news too.

Excellent Analysis!

Basically, UPST is still one of my highest conviction holdings. Even considering conventional P/E valuations, this stock is a buy now.

The only real problem I see, and it is central to all our discussions and projections, is that we simply don’t have enough history or at least not enough stable/consistent history to predict the future with a reasonable level of certainty.

Yes, it’s not SaaS with its recurring revenue as we all know. But more to the point, just looking at the last 2 quarters, it’s impossible to predict the next quarter, or year, with any level of confidence. All analysis points towards a fairly glorious future, BUT it is the very very low level of confidence that is hampering estimates. The company is simply too young, the business model at a too early stage of its development.

But I’m convinced this is a fantastic entry point for anyone who still has funds to invest and isn’t already heavily invested on UPST (I’m out on both counts unfortunately).