Description

Thryv is a growing, profitable SaaS business inside a melting ice-cube, and for once the melting ice cube is actually a benefit to the profitable business. I’m sure we’ve all seen so many melting ice-cube companies trying to pivot that we automatically roll our eyes at the thought, I know I do. However, Thryv actually has major synergy and is achieving a very low customer acquisition cost for their rapidly growing SaaS product thanks to their legacy melting ice-cube.

Business Overview

The SaaS business

Just as HubSpot is the low-cost disruptor to SalesForce [1], Thryv is the low-cost disruptor to HubSpot. Thryv sells software for running a small to medium sized business and makes it very simple so that the business owner doesn’t need any computer expertise. Thryv has a lower cost offering that suits the very, very many small and medium businesses that are finally moving online in large numbers and their secondary business gives them a much lower Customer Acquisition Cost (CAC). Their offering is building more feature sets and moving up into HubSpot’s territory while building from a lower cost base (the Clayton Christianson classical definition of market disruption).

The Directory Business

This is the melting ice cube. Thyv, the company, is directly descended from Dex Media, owner of many old Yellow Pages. These days phone directories include online listings but are still a fading business used most by the 55-and-older demographic. Thryv has been going around with their knowledge of the directory business and acquiring the remaining directory companies. I’ve been watching their technique for years. They were basically borrowing money and buying directories at what amounted to full net present value - the cash streams would be expected to pay for the buyout debt and that was it. Nobody would bid higher than that so they could pick any target they chose. The upside was that they expected to be able to sell Thryv SaaS to some of the customers they had directory relationships with. At first that was theoretical and I was skeptical, then it became a success. Their current new-customer mix for SaaS is about 1/3rd directory customers, 1/3rd referrals from directory customers, and 1/3rd more conventional channels. 

Bear Argument: If directory cash flow decreases faster than expected the melting ice cube could drown the SaaS business.

This was my first take on the company a couple years ago. I’ve seen way too many melting ice cubes with management making bold claims about being able to pivot. I still have my first excel model of the company where I expected the directory business to basically cancel the debt at best. Then Thryv SaaS started getting lots of customers from the directory business. After that, they started to make the directory sales staff available to sell Thryv SaaS by slowly lengthening the cycle time of the directories. That was very clever - lengthening the directory cycles from 12 to 15 months effectively allows the company to move 25% of the directory sales staff time over to selling SaaS while also reducing directory business costs with very fine control and cost efficiency. Then they moved from 15 months to 18 months, and the clients still pay by the month for directory services so no cash flow is lost. Finally, they managed to slow the descent of the directory cash flow and pay off debt faster than expected from business earnings. My most recent estimate is that the melting ice cube has a significant NPV even without counting the benefit to SaaS sales.

The overall benefit of the directory business has been a big boost to the core engine of the business. It provides ready sales staff in whatever proportion needed, and provides about 2/3rds of the SaaS customers at extremely low Customer Acquisition Cost.

The Narrative and why the stock is cheap

We all know what the market wants to hear about SaaS: margins and growth. Thryv SaaS margins are great, as is typical for SaaS. But a simple look at their customer count would leave one under the impression they are not seeing traction. Back in 2018-2019 they had a base price for their main tier of $199/month. Management announced they were shifting to a base price of $349/month and that they were going to let go of old users who didn’t move to the new price plan. This was followed by 12 quarters of management trying to explain that they were shifting user bases while the top line announced total customer count decreased or flatlined. This is a chart I managed to build with some linear algebra: 

Figure 2, Subscriber population:  The gray line is the number of total customers reported each quarter. As you can imagine, that decrease and apparent stagnation was not popular with the market. The blue and orange numbers were not reported (but extractable with sufficient effort). I’ve been watching the orange line. Once it caught up to the gray line I wanted to see the gray line start to rise again, and it did. One could speculate that the true customer growth is closer to the orange line slope, but some of that will be conversion from the blue line too, so maybe going forward we might expect growth between the recent gray line and the orange trend line. 

Mathematically speaking, the raw orange trend line predicts more customers in 2022 Q3 than were actually reported but recall that this is “customers at $349/month.” If we compare predicted revenue we get a closer fit because ARPU is also rising. This means management is making trades between per-user revenue and raw user growth which appear to be pretty efficient

Here is another way of looking at it:

Figure 3, ARPU (monthly): The horizontal dotted lines show old and new base prices. Thryv is now building on a base of users at the upper pricing and raising ARPU with feature add ons.

To be honest the thing that most impressed me about the transition shown above is that they told us they would do it, discussed it along the way, and then accomplished it. I actively track management predictions and promises versus results and this was an excellent example of management deciding to do something that wouldn’t be rewarded by the market, and doing it anyway for long term improvement. Some of the conference calls along the way were kind of humorous, with management making wry comments to each other about how bad the subscriber numbers look despite the underlying structure of the company improving.

Bear Argument: Customer count has hardly budged in years. No growth.

Basically, I spent the last two years specifically trying to prove or disprove this bear thesis by tracking the customer populations as plotted above. With the recent complete conversion of the customer population to the new higher price - combined with growth in the Q3 2022 quarter - we can finally say that this bear argument does not hold. The apparent stagnation was a bold and deliberately executed step taken by management and we have extracted the data to see that they have a healthy and growing customer base who is willing to pay more than before.

Competition (and Moat)

There is a long list of companies trying to sell software to run your business. My notes go pretty deep into at least 27 of them. Some target mostly the email side and don’t do quotes or invoicing or billing, others focus on B2B or specific industrial verticals. One can keep coming back to the fundamental question: why should we expect THRY to be able to continue to exist and grow in the market? This question is especially interesting because this is a market where a lot of the players are not profitable - they spend more on customer acquisition than they have demonstrated they can generate in customer long term value. Hubspot (HUBS) is a poster child for this behavior and talking to various folks suggests that their spending on R&D includes a lot of tasks that others might classify as Field Application Engineering (FAE) - making specific changes and adaptations to keep and support customers - that cannot be eliminated without losing customers.

The moat and secret sauce for Thryv is their old phone directories business and how it gives them a large and ridiculously cheap pool of customers. Once a SaaS in these markets has profitable scale it gets much harder to replace them as any competitor will be coming in at less-than-sustainable scale while facing pressure on both CAC and pricing from the incumbent. So how did THRY disrupt under HUBS and reach a sustainable scale? A big part is a combination of the reduced CAC from getting customers through their directories business and the reduced growth costs they get from slowly transitioning directory business sales and support staff over to SaaS. The reduced CAC has helped them grow while remaining profitable. As long as they keep operating in this profitable regime, they will be very hard to unseat because they can always compete effectively with would-be competitors. As they build scale they will eventually run out of convertible directories customers, but by then the math suggests they will be quite astoundingly profitable.

With the $479/month service level Thryv designs and maintains a business’s website using thryv back end support. For example, you can see a website provided by Thryv to a small business here: https://www.bigbluewrench.com/

This is a customer who used to be in the phone book and now has a website that makes appointments and allows customers to log in to see their work schedule and pay their bills. The business user just uses an app on their phone (or web) to keep track of their schedule, click to send invoices, and receive money all automatically. If you dig around in the code for that site you’ll see the use of a number of back end services calling directly to the thryv service infrastructure:

At one point I was using web code search services to watch sites that use such calls and watched them rise significantly in the last year, then it looks like Thryv changed their coding tools so now the calls are not as uniform so my data history is no longer apples-to-apples but their presence still seems to be growing.

How good is the directory→ SaaS pipeline? We spoke to one Vivial directories customer who was initially not very interested in the Thryv SaaS product. After 8 months and 3 calls from Thryv (from the same exact person they had dealt with for years to be in the Vivial directory) the customer business had a full SaaS suite and website provided by Thryv. Now most of their booking, billing, and pay is through the Thryv SaaS tools and the customer never had to become any more tech savvy than learning to use a new app on their phone (and the reps will gladly walk you through that too).

Debt

In addition to the misunderstood customer growth of the last couple years, another thing that has folks worried about THRY is their debt. This debt was taken on to fund purchase of their directories for the directories business.

Bear Argument: So much debt can sink the company

As of the 9/30/2022 quarterlies the combination of short and long term debt is about $500mm, if we root around in the contingent liabilities of various contracts we can find maybe $40mm where it’s a little unclear if the company is consolidating those for reporting or not, so let’s call that $540mm. There’s also a pension liability (the plan is frozen) that eats a chunk of cash each year as they move it to fully funded. So this is worth digging into.

The debt concern is actually a very reasonable concern and possibly a contributing reason THRY stock is down so much since the fed started raising rates. As a counterpoint, I’ve had a LOT of success buying good businesses in debt and watching them pay down their debt, increasing both the effective value of the shares and the optimism of investors towards them. 

The main chunk of debt is a term loan with interest on a base rate (highest of prime, fed funds, or LIBOR one month plus 1%) added to 7.5%. When rates were lower and the stock was higher this resulted in 8.75% interest. Right now I calculate they are around 11.75% interest as of January 2023. They have a mandatory $17.5mm principal installment every quarter, although they tend to pay more. Here is a plot of their debt and repayments over time:

Figure 4: Debt and debt payments versus time

When modeling the fate of the debt we have to consider what the future Cash Flow of the melting ice cube (the directories business) looks like. Based on historical performance we assume revenue shrinking by 22% per year (retaining 78% of the previous year’s revenue) and a similar decrease in maintenance capex. Figuring out the remaining costs is harder. I worked with two EBITDA models. The first just assumes EBITDA shrinks with the business (retaining 78% each succeeding year). This is probably overly optimistic considering there are likely some fixed costs that won’t scale down smoothly with revenue - I refer to this as the OVERLY OPTIMISTIC MODEL. The second model takes the historic total EBITDA to Directories Revenue and linearizes it:

Figure 5: Modeling effects of scale on directories business profitability to include scale loss in projections of whether the directories business can pay off the debt.

I refer to this second model as the OVERLY PESSIMISTIC MODEL because the EBITDA data is not only for the directories (that isn’t disclosed separately) so the directories here are paying the net costs of the SaaS division before we get the profits and those SaaS costs have increased in recent quarters as the directories revenue goes down, making the slope here worse. (Note that the slow gains in cash flow of the SaaS is probably what’s causing the noisy behavior in the left side of the plot). Additionally, the directories staff are being slowly transitioned into supporting SaaS and the accounting for that is unclear but should favor lowering directory costs.

Note that the debt has an “excess cash flow” covenant so management must pay down the debt with any excess cash flow. Meanwhile the increased interest rate going into 2023 is likely to cost Thryv about an extra $10mm in interest. Management know this - during their Q3 2022 call their CFO commented:

“As you can see by our debt repayment this quarter, we're focused -- just, well, we have always been focused. You know my 2 commandments about paying down debt. So we're focused like a laser beam, paying down debt all the time, unless where you got some outstanding opportunity, an acquisition that comes our way. So -- and given the interest rate environment, it's in our best interest and our shareholders that we pay all the debt down we can. So that's been our focus.”

Therefore, and to keep things simple, I just model the cash flow of the directories business going to pay down debt.

Interest rate at 8.75%

In my original models 12-18 months ago it was fairly clear that the acquired directory businesses would pay off the debt that was used to acquire it. Modeling the directory business in this case winds up with pay off of the term loan in early 2026 for the OVERLY OPTIMISTIC MODEL and with remaining debt down to $100mm (a reasonable long term amount) for the OVERLY PESSIMISTIC MODEL. 

Interest rate at 11.75%

In the OVERLY OPTIMISTIC MODEL we see debt paid off completely in 2026. In the OVERLY PESSIMISTIC MODEL we see the term load down to about half its current level in 2025 and getting down to $100mm around 2027 although it should be noted that this particular model uses the 11.75% debt rate for that entire time, which seems like a rather dark projection.

Note that there is a lot of minutia and side notes here. The loan is likely to get refinanced before the ends of most of these projections. There are other expenses like the (frozen) pension obligation and long term facilities leases. There is also a whole other business (the SaaS side) which has revenues, EBITDA that is positive in established geographies, and so on. As a practical matter management has not in previous years put all of this cash flow towards paying down debt as they have instead used it to acquire and build for the SaaS.

Basically this is a safety check to see if the debt is likely to sink the company or if we can count on the good business to eventually rise out of the debt pile. I’ve run a lot of interest rate scenarios and under most non-apocalyptic scenarios the debt should be paid off by the melting ice cube business just fine. Any refinance would be a big gain as well - the debt they already have was acquired after a bankruptcy and during a time when the future of the company was quite uncertain.

Valuation

As a simple approximation based on the models above I value the debt and directories business together at a net value of zero. The directories business is worth about as much as the debt, even though management will probably continue to pull some money out for now to build the SaaS business and as a result wind up with some ongoing debt in the future.

The current SaaS growth rate is running at 26% YOY but even my best case valuation only goes with 4 years of 20% growth followed by a step down to 8% growth even though they are in a rapidly growing market. My low growth rate estimate is only 11% growth, but for a few more years than the medium growth model. I’ve used a blended 8.75% long term discount rate but the results are still more than 2x gain with a 10% discount rate. 

For cash flow on the SaaS, based on mature geographies being EBITDA positive, I put an approximate 20% cash conversion estimate on the SaaS revenue. From there, it is a pretty standard DCF.

The valuation conclusion? A 300% to 400%  valuation versus current share price. I wouldn’t be surprised to see an $80 share price in the not too distant future.

Now, there has been a fair amount of philosophical talk about how to properly do a DCF for a growing company that puts cash flow back into growth, which is probably how things will play out for THRY for at least the near future. I fully agree that cash flow put back into the company for growth is accounted for twice if we count it as cash flow when earned and then also enjoy the fruits of the growth it buys in the future. So while the above remains the “standard” way of calculating DCF for most people, I also ran versions where I give the company zero value for the next seven years of cash flow. Fortunately that still leaves us with about a 260% to 300% valuation compared to the current share price - most of the value is in the out years after some near term growth..

[1] In the classic Clayton Christensen sense - coming in at the bottom of the market with a cheaper product with the ability to build upwards into larger existing markets

Catalyst

Two main things are going to push the share price higher over the coming years. First is debt repayment, as the company pays down debt and pension obligations that enterprise value accrues to shareholders. Second is rising EBITDA from SaaS. The SaaS is already EBITDA positive in established geographies, but not overall as they are still putting cash flow back into adding geographic coverage. A debt refinance on better terms (current term loan is base rate + 7.5% , while revolver is base rate + 2%) would be a sudden jump up in financial situation. A large shareholder specializing in distressed companies (from all the way back at the bankruptcy reorg) has also recently finally finished unloading shares (many of which I bought), so that is one force towards more positive price action.