Upstart Concentration Numbers

Investment Analysis Clubs Saul’s Investing Discussions
1

Feel free to double-check my numbers but when you break down the concentration risk for Upstart, there are some interesting findings.

Let’s start with Cross River Bank (CRB).

Here are the last 3 blurbs from UPST’s 10-Q’s:

From Q3:

In the nine months ended September 30, 2020 and 2021, fees received from CRB accounted for 65% and 59%, respectively, of our total revenue

From Q2:

In the six months ended June 30, 2020 and 2021, fees received from CRB accounted for 70% and 62%, respectively, of our total revenue

From Q1:

In the three months ended March 31, 2020 and 2021, fees received from CRB accounted for 79% and 60%, respectively, of our total revenue.

If you break out those numbers, you get something like this:

CRB Revs

	Q1	Q2	Q3					
2020	51	6	38	
2021	73	123	125					
				
				
				
Non-CRB Revs
	
	Q1	Q2	Q3				
2020	13	11	27	
2021	48	71	103

And here are the sequential growth rates for these:

CRB

	Q1	Q2	Q3					
2020		-87%	499%	
2021		49%	2%					
				
				
				
Non-CRB
	
	Q1	Q2	Q3				
2020		-18%	146%	
2021		47%	44%

So what we find is that CRB is slowing down quite a bit but the other banking partners are really ramping up. That’s exactly what we want to see.

If we break out the concentration risk by quarter, we get something like this (note that this is by quarter and the 10-Q blurbs are 3, 6, and 9 month periods respectively).

CRB concentration
			
	Q1	Q2      Q3
2020	79%	37%	59%
2021	60%	63%	55%

So we see that CRB accounted for 55% of revenue, down from 63% just last quarter.

And now for the Credit Karma numbers.

From Q3:

For example, the nine months ended September 30, 2020 and 2021, 52% and 44%, respectively, of loan originations were derived from traffic from Credit Karma

From Q2:

For example, for each of the six months ended June 30, 2020 and 2021, 49% of loan originations were derived from traffic from Credit Karma

From Q1:

For example, for the three months ended March 31, 2020 and 2021, 48% and 53%, respectively, of loan originations were derived from traffic from Credit Karma.

Note that these are originations so I used the number of loans rather than revenue.

CK Originations	
			
	Q1	Q2	Q3	
2020	40,423	6,661	44,947	
2021	89,968	133,773	136,793					
				
				
				
Non-CK Originations

	Q1	Q2	Q3					
2020	43,791	5,215	35,946	
2021	79,783	153,091	225,988

And then the sequential growth rates would look like this:

CK Originations	
			
	Q1	Q2	Q3	
2020		-84%	575%	
2021		49%	2%					
				
				
				
Non-CK Originations

	Q1	Q2	Q3					
2020		-88%	589%	
2021		92%	49%

Once again we see that Upstart is doing a really good job of decreasing their concentration risk. These Credit Karma numbers are very impressive to me because it’s tough to create completely new marketing channels.

And the quarterly concentration numbers look like this:

CK as a % of originations
	
	Q1	Q2	Q3				
2020	48%	56%	56%	
2021	53%	47%	38%

So Credit Karma concentration is down from 53% two quarters ago to 38% now.

All in all, Upstart is executing very well in terms of decreasing their concentration risk. It’s also helpful to see how fast they are growing outside of CRB and CK. This gives me more confidence they will continue to grow.

Best, Fish

123 Likes
2

Cool perspective on the numbers! I think it shows how the dynamics of growth works:

  • once a new partner signs up, it takes about 9 mths to ramp up fully. In the 4th quarter after sign-up the numbers stabilize.
  • in the upcoming 2020 Q4 I would thus expect to see further growth from the recently gained bank partners.

The flipside of this model is that UPST would need to sign up new bank partners for continued growth. I would expect thus expect a more lumpy growth path: If they win many and large banks, there are three quarters of fast growth, if they win only a few smaller banks, the growth is slower.

I have no idea how large the recently added bank partners are but I think this detemines the growth in the next two quarters. But it has a higher level implication: their growth depends less on their AI model but rather on their ability to sign new bank partners.

The long term thesis remains of course that UPST with its superior AI approach is attractive enough to (i) win plenty of bank partners (and with them their customers) - and (ii) enter new market segments which would start similar growth paths.

To me, this even has a land-and-expand element to it. I would expect that a partner bank (once landed) is easier to cinvince to try new market segments (expand). But ultimately it depends on the sales organisation to win bank partners.

LNS

16 Likes