An attempt at an economic model for SaaS comps

Investment Analysis Clubs Saul’s Investing Discussions
1

I decided to make a concrete dollar model to try to make the SaaS economic model more understandable (to me as well, because it’s hard to get my mind around it and actually grasp what they are doing). I’m guessing at the figures and will be glad to have others make suggestions, but my figures do seem to come out pretty realistic (in the ball park, anyway). Here goes:

Let’s say our SaaS company, ABC, has $100 million in annual revenue. It’s growing revenue at 60% and will thus grow revenue by $60 million this year. For simplicity sake let’s say it adds $15 million each quarter.

If ABC was selling refrigerators it would have to cost them a lot less than $15 million in S&M each quarter for them to be profitable, but ABC is selling a complicated set of programs that have to be integrated into the customer company’s business with no resulting problems, and this takes a lot longer and a lot more effort than selling refrigerators. Let’s say that ABC will be happy if the first year’s lease covers the cost of S&M… a “one-year payback,” as they will have the customer for many years (which is the key to their model).

Let’s say there is a one-year contract paid in advance. (If someone wants to do the model with a two-year contract that would be nice, or altering the payback period, or whatever)

Okay, we had decided that they sign a quarter of the new customers each quarter:
The first quarter new customers pay for 12 months but on average only 10½ months are recognized as revenue in 2018
The second quarter new customers pay for 12 months but on average only 7½ months are recognized as revenue in 2018
The third quarter new customers pay for 12 months but on average only 4½ months are recognized as revenue in 2018
The fourth quarter new customers pay for 12 months but on average only 1½ months are recognized as revenue in 2018

So 24 months of new customer payments get recognized, out of 48 months actually paid in 2018. (On average, it’s as if they all signed up in the exact middle of the year, all paid one year’s lease, and all only had six months of that lease recognized as revenue in 2018).

We said that ABC would be happy if the first year’s lease covers the cost of S&M (one-year payback). So ABC paid out the equivalent of 48 months of new contracts as S&M expense in 2018, but only 24 months got recognized. Are you with me so far?

What does that mean in dollars? Let’s go back to see:

ABC has $100 million in annual revenue. It’s growing revenue at 60% and will thus grow revenue by $60 million this year. But they have a dollar-based net retention rate of 120%, so they only are recognizing $40 million in new-client revenue. And we just figured out that they paid $80 million in sales and marketing expense in 2018 to get the new $40 million that was recognized. (This doesn’t worry them in the slightest because the rest of those first year contracts will be recognized during the next year with no new S&M expense. And they already have the money in the bank).

Let’s go on. That $20 million that they got in expanded sales from existing customers probably cost them something too, but these are established customers, so it doesn’t cost as much to sell them. Let’s say just it costs just a quarter as much. The $20 million recognized came in from added contracts sold spread out over the course of the entire year, so it means there were $40 million in actual added annual contracts, of which only $20 million was recognized as revenue in 2018, and since these were only a quarter as expensive as the new contracts, the $40 million cost just $10 million in sales and marketing expense.

Finally we have the $100 million existing contracts that were renewed during the year. For simplicity, lets say they only cost one-tenth as much as selling a new contract so they added another $10 million in sales and marketing expense.

Thus for 2018, we’ll have $100 million in sales and marketing total expense (80 + 10 + 10), and $160 million in recognized revenue. Thus S&M expense came to 62.5% of recognized revenue, which is pretty representative of a lot of our SaaS businesses. When you add in R&D and G&A expense, that’s probably enough to give them a small “loss” on the year.

But should we worry about that loss? They re-signed $100 million of existing customers at a total cost of $10 million. They upsold $40 million per year, at a cost of $10 million total, and they signed $80 million per year in new business for $80 million, but those new customers will renew next year probably for $96 million at a much lower S&M expense (120% dollar-based net retention).

I recognize that there are flaws in this simplified calculation of course, lots of them that I can see myself, but as a birds-eye overall view, and an attempt to make a complicated situation understandable, I hope it helps. If someone else can make a better example it would be wonderful! I’m worn out.

Best,

Saul

50 Likes
2

Hi Saul.

It’s earnings season, and I can’t give this as much attention as I’d like to right now (actually, it seems like a “fun problem”).

To me, it sounds as if the crux of your GAAP concern – in this particular case, I realize it isn’t your sole GAAP concern – is that sales & marketing (S&M) is expensed in year one, when its useful life should be longer. How much longer do you think it should be? Looked at another way, I’m sure that the commissions paid to the salespeople somewhat reflect the fact that revenues are likely to be recurring, rather than just reflecting the first year’s cash flow. Competition for the best salespeople would require bidding the commission up towards some percentage of the economic value of the sale to the company, right? I mentioned Zayo in a prior post, and I know they do something very much like that, basing the commission on the expected future cash flows, not just the first year’s cash flow. They want their salespeople properly thinking like owners, and good incentives support that. Do you think three years is right? Five…? I know that “perpetuity” is too long.

Another question that might help with setting up a spreadsheet to work through the recognition of revenue and expense… Do you want to work with 60% revenue growth and 120% retention as a long-term assumption, or do you feel as if either number should decline over time?

I can’t swear I’ll attack this problem, but I do find it intriguing.

Thanks and best wishes,
TMFDatabaseBob (long: ZAYO)
See my holdings here: http://my.fool.com/profile/TMFDatabasebob/info.aspx
Peace on Earth

2 Likes
3

Bob, It’s all just an approximation. Pick your own numbers and see how it runs. You have to assume though that a SaaS customer company is signed on forever, for all intents and purposes.
best,
Saul

4

It seems to me that there are two things mushed together here.

If one pays commission on the initial sale, yes, that could be considered as producing revenue over the life of the contract, but a lot depends on the nature of the commission arrangement. If commission is paid on renewals (perhaps at a lower rate), then really each commission only impacts one year and I think could easily be ignored in this model. Likewise, marketing is an on-going activity.

Note that, if I am selling a software package which goes for $500,000 in a conventional sale and I price it at say $125,000 per year in a SaaS model, the sales people will be really annoyed if they only get a commission on the first year’s sale instead of the $500,000.

But, in Saul’s description, it sounded like you were also recognizing some start up cost for getting the new account going, what I would think of as implementation. There is also some on-going cost in taking care of the account over time, what I would call maintenance. But, depending on the nature of the software, the implementation effort can be substantial. One might wish to amortize that implementation cost since it is unlikely to recur, unless there is purchase of new models or the like.

1 Like
5

Note that, if I am selling a software package which goes for $500,000 in a conventional sale and I price it at say $125,000 per year in a SaaS model, the sales people will be really annoyed if they only get a commission on the first year’s sale instead of the $500,000.

You just change the commission plan to work on booking value and not revenue realized. If I give a sales guy a 100k salary (base + variable), I just make the target 500K booking instead of $125k revenue. That way the incentive is to drive higher booking, cash paid up front.

It is hard to track revenue to salespeople for service contracts done in the past. What if the salesperson changes after year 1, is the new salesperson suppose to get credit for the 125K earned in year 2? So just give them booking targets,

2 Likes
6

JDC,
So I don’t know a lot about how s/w sales works and how commissions are paid, but your comment:
You just change the commission plan to work on booking value and not revenue realized. gave me pause.

What if the revenue is never realized? The company gets acquired and is forced to adopt the s/w standards of the acquiring company (happens with just about every M&A). The company goes out of business. The company decides after a year they want to use different s/w, etc. There are myriad reasons that a company may discontinue a subscription.

Yes, it is disruptive, but the disruption is not always a deterrent (especially with M&A situations). In my career I worked on more than one of these (although it was before the days of SaaS). IMO customer lock-in in a fallacy. No customer changes out s/w that is integrated with their operations wantonly, but reluctance is a far cry from never.

Anyway, what happens to the 4 years of commission already paid out if the customer bails after one year? Do you claw it back from the salesman? Can’t imagine that’s gonna sit well with anyone on the sales team right on up to the director who got a bonus on bookings rather than revenue.

7

What if the revenue is never realized? The company gets acquired and is forced to adopt the s/w standards of the acquiring company (happens with just about every M&A). The company goes out of business. The company decides after a year they want to use different s/w, etc. There are myriad reasons that a company may discontinue a subscription.

Yes, it is disruptive, but the disruption is not always a deterrent (especially with M&A situations). In my career I worked on more than one of these (although it was before the days of SaaS). IMO customer lock-in in a fallacy. No customer changes out s/w that is integrated with their operations wantonly, but reluctance is a far cry from never.

Anyway, what happens to the 4 years of commission already paid out if the customer bails after one year? Do you claw it back from the salesman? Can’t imagine that’s gonna sit well with anyone on the sales team right on up to the director who got a bonus on bookings rather than revenue.

Basically, yes, you claw back from the Sales buy just like if you were to sell revenue on HW and the customer later returned. It doesn’t always sit well with the sales guy but it is defined in the sales program that they are measured on. I haven’t worked in a SAAS model but I did run the sales compensation team for a company doing HW and service contracts. Part of the incentive on services would be to retire quota on the full value of the service booking if the customer was paying upfront for the full term of the contract.

But you balance it out by giving them larger quotas then if they were measured on revenue as you would expect the 3 year service contract to be much higher.

The other incentive of doing it this way is that the sales team concentrated more on booking sales and not getting too involved in internal process issues on how to get revenue recognized. But when we had their compensation on revenue, they might spend too much time pushing the delivery team or other function on completing tasks so that revenue could be recognized. We wanted them customer focused and not operational focused

1 Like
8

isn’t the key for the SaaS model, is how to value the subscriber of the service?

25iq.com/2017/07/15/amazon-prime-and-other-subscription-busi…

9

It is hard to track revenue to salespeople for service contracts done in the past.

No, because the salesperson earning the commission in year 2 is the person handling the account in year 2, whether the same or not. They get the renewal signed, audit user counts, sell add-ons, etc. One might make the commission a higher percentage in year 1 to reflect the draw of new business, but the account requires on-going handling in subsequent years.

1 Like
10

No, because the salesperson earning the commission in year 2 is the person handling the account in year 2, whether the same or not. They get the renewal signed, audit user counts, sell add-ons, etc. One might make the commission a higher percentage in year 1 to reflect the draw of new business, but the account requires on-going handling in subsequent years.

When the order is booked it usually just becomes an item on the GL that then can recognized by an amortize table. The month those contracts are booked, they are easy enough to track and assign. In year 2, Sales leaders have changed their plans, maybe redesigned territories, new incentives have been added to increase a focus on services etc. Now when we recognized service revenue on year 2, it is just a GL entry to move deferred revenue to revenue that is made up of 1000’s of contracts that almost nobody is going to try to figure out where the credit should go.

I have seen it done at 3 companies, and none tried to recognize credit on the revenue of the contract, just the booking due to how difficult it is to track. Either manually or using a tool like Xactly given that usually you only get 1 or 2 people to calculate and pay commissions for 300_ sales ppl.

11

Saul,
…as they will have the customer for many years (which is the key to their model).

It is my considered opinion that you put too much weight on the longevity of subscriptions.

During my 30 year tenure in IT at a very large corporation I worked on two acquisitions and one divestiture. I also worked on several projects that were not specifically M&A type activities, but placed me in a direct working relationship with consultants from firms like Accenture, PWC and others with units that specialized in these type activities. The very fact that these firms had groups that specialized in this type of activity is a testament to how frequently they occur and the direct impact on IT operations in general and s/w strategy specifically.

From my experience, I think it’s safe to say that the vast majority of M&A transactions result in the acquired firm having their entire IT stack disrupted in order to be compliant and compatible with the acquiring firm, the exceptions being Berkshire type acquisitions that pretty much don’t interfere with the operations of an acquired firm.

This is not the only reason for software subscription to likely be terminated. Companies go out of business. Companies get disenchanted with the a given s/w package as the ROI is not realized. Companies determine that there’s a better, more economical alternative. Companies at times cut costs, IT is a cost center for most companies so they often bear a disproportionate burden of these mandates. The s/w vendor fails to address bugs and other problems with their s/w. The s/w vendor fails to deliver on promised enhancements and upgrades. The s/w vendor gets acquired by a competitor which was more focused on acquiring a customer list than perpetuating a s/w package. The s/w vendor alters their business model which is viewed by a segment of their customers as detrimental to the relationship. And so forth . . .

I’ve personally witnessed almost every single item in the list (with the exception of the non-specific "And so forth . . .).

Of course there’s a cost associated with a transition from one s/w package to another. But the cost of transition, in and of itself is not a major inhibition under most circumstances. Every project I ever worked on looked at ROI, payback period and a host of other financial pro forma analysis over a five year window (sometimes longer) in order to determine a course of action. On at least one occasion, a multi-year project of unparalleled proportions that completely upended the existing engineering/manufacturing information pipeline was undertaken because the executive in charge determined that it was the right thing to do. As an aside, I personally attended a couple of meetings with this executive. He was by far the most brilliant business person I have ever known.

I’m not saying that companies incur the cost of transition wantonly. Often the s/w cost is the least of many considerations; there’s training costs, file conversions, procedural revisions, testing and integration costs, internal resistance to change, etc. There are a lot of considerations. A pervasive product like a standard DBMS, PDM, ERP, CRM or similar product is a very expensive and disruptive change. A niche product like AYX or even OKTA might be swapped out without much operational impact.

And I’m not saying you’re wrong, but maybe just a bit too sanguine about the strength of the SaaS subscription model. I really don’t see this customer glue to be any stronger than an purchased license with follow-on maintenance contracts. Each past year is a sunk cost. Each coming year is a new year with new competitive forces and at times, new decision makers.

39 Likes
12

JDC, it is a very different issue if the original contract is for three years with unlikely cancellation provisions versus the original contract being for one year with expected renewal. I suppose the former happens, but my experience is that the latter is much more common. In the former case, there may be little or nothing to do in year two or three, but I think this is likely to be uncommon in SaaS companies since one of the whole ideas of SaaS is being able to dynamically size to changing business requirements. This implies a need for significant customer contact to insure renewal, assess usage levels, explore new functionality, etc. If that work is done by the original salesperson, then fine, but it may well get transferred to someone else for many, many reasons. I can’t imagine paying commission up front for 5 years worth of expected renewals. I can see keeping track of the original salesperson and the current salesperson and giving them each a commission on renewals and add-on sales.

As a point of reference, I once had a customer where the original sale was about $700K, hardware and software and initial mods, but because I was able to respond very economically to enhancement requests their 5 year spend was well over $2M. And, that wasn’t even SaaS, where the initial sale would have been spread out. Commission structures have to be fair or people get unhappy.

1 Like
13

Brittlerock, you are bang on. One of my best customers had been a customer of another package for years and years, but they were quite unhappy because any modifications were very slow and very, very expensive, so they decided to look around. I sold them based on their belief that I could do much better. Not only was the original installation a major improvement on what they had, but the low cost of enhancements had them deciding left and right to make changes and additions to better fit their operations. Their CFO used to go to conventions and brag about what they could accomplish … but they also spent a lot of money with me, growing 10X in revenues while they were my customer with zero addition to HQ operations staff and maybe 30% more in the warehouse because of increased efficiency. Then, of course, they were bought by a much bigger company whose software wasn’t nearly as good, but you know how it goes.

2 Likes
14

it is a very different issue if the original contract is for three years with unlikely cancellation provisions versus the original contract being for one year with expected renewal.

Those are completely different. In your case, of courses the sales should get credit each year as the renewal will trigger a new PO and booking. I would still suggest that you give the credit on booking value for the 1 year contract then. If a 100k order is booked in Q1, you don’t split it over 4 quarters, better to just have the rep driven on bookings. If the order cancels in Q2, claw the full value back.

Likely SAAS contracts are only 1 year and always need to be renewed in which case that would make sense.

15

First Off,I recc’d your original post. I have been thinking a bit about the SAAS market and the rather quite remarkable run almost the entire sector has been on for some time. I have also come to the conclusion that the market was mispricing these companies because of the very reasons you state. Namely that the short term expenses and the long term earnings are somewhat mismatched in the model. Further your attempt at trying explain the very mismatch with an example is quite good and enlightening to the very thing i had been thinking about.

I guess the big question I have is what is the proper valuation method? It makes me a little uncomfortable to just say the valuation is not important and that only growth rates matter. I agree that valuing on the traditional measurement of earnings in some manner doesn’t work here as many companies are not showing earnings, but that doesn’t stop the need for valuing. P/S may be good, but there needs to be some measuring stick. Otherwise we can easily fall into the 2000 market type of trap where the stock price going up is the only thing you need to worry about.
Despite how good and simple your examples are, you don’t in any way try to quantify what is a fair price. When I first started following this board you were always looking at growth rates relative to PE. Now i dont seem to get any feeling of a relative valuation methodology.

Far be it from me to tell you how to pick stocks, that clearly wouldn’t make sense. It just surprises me you don’t seem to holding these stocks to any type of valuation comparison (other than who is growing faster or getting the highest retention rate). When is the price too high? Is SHOP too high at $125? Clearly you dont think so (and i dont either). But what about $200, $250 or even $500. No matter the growth rate, some price is too high as the growth rate cannot stay forever (or at least the chance of that is pretty slim).

No criticism here, just interested in your thinking.

Randy
Long SHOP

19 Likes
16

I too found Saul’s posts about valuation etc. (40408, 40439, 40464) extremely interesting and helpful - thank you Saul for all that work and the thinking behind it.

I did feel throughout that, having seen more severe market downturns than I would like, my big takeaway is (italics): everything changes. Previous certainties are trampled with all the shock of the new invention of the tank across rural Europe in the early C20.

Broadly agreed with brittlerock’s post and add only that it is the best, above all the favorite of peers, which counts at board level; decision then simply delegated for implementation. The receiving company will make it easy.

Agreed with Randy that some method of valuation makes me a happier investor. Can tolerate missing out, or delegating to a fund - though with a certain scepticism about their actual abilities given the momentum of the last few years. Of course, there is no fund which replicates this board!

The essential features of preparedness are:

Recognise that high valuation ratios are dangerous and multiple reversion to the mean is normal, which in full bear markets overshoots to the downside before recovering to the mean, which may be much less than now.

Look for low or no debt (if cap. halves, debt doubles: how does the company look then?).

Look for pristine balance sheets and pricing power.

4 Likes
17

Happy Anniversary streina!

18

I’m not saying that companies incur the cost of transition wantonly. Often the s/w cost is the least of many considerations; there’s training costs, file conversions, procedural revisions, testing and integration costs, internal resistance to change, etc. There are a lot of considerations… A niche product like AYX or even OKTA might be swapped out without much operational impact.

Hi Brittlerock,
I think there are some things you are not taking into consideration, aside from all the hassle of switching.

First, Alteryx grew revenue at 56.5% last year. Okta grew revenue at a rather incredible 65%. That means people really LIKE what they are selling, and they will have no reason to switch unless something much, MUCH better comes along.

Second, Alteryx had a rather incredible Dollar-based Net Retention Rate of 133% last year. Okta’s was over 120%. That means people really LIKE what they are selling, and they will have no reason to switch unless something much, MUCH better comes along.

And here was Pigskin’s report from the field last month:

OKTA is awesome. The company I work for started using these guys a few months back and it has made my life soooooo much better. Before that I had to remember tons of passwords as we use a lot of different programs. I was constantly signing in and out of software. I have no idea how this translates to money saved by my company. I am more productive and now I am not calling IT about getting something unlocked because I forgot the password. There are clearly some savings involved. OKTA has even come up in a few meetings where people are giving accolades to the move, so it has created quite the impression… From a minion’s perspective in a large corporation these guys get two thumbs up.

We’ve had similar reports on Alteryx in the past.

Do you really think a customer company is going to switch out of companies like this to save a couple of dollars, or on a whim. Their employees would rebel! People really LIKE what these companies are selling, and they will have no reason to switch unless something much, MUCH better comes along.

That’s just my non-techie, “common sense” point of view.

Best,

Saul

16 Likes
19

Far be it from me to tell you how to pick stocks, that clearly wouldn’t make sense. It just surprises me you don’t seem to holding these stocks to any type of valuation comparison (other than who is growing faster or getting the highest retention rate). When is the price too high? Is SHOP too high at $125? Clearly you dont think so (and i dont either). But what about $200, $250 or even $500. No matter the growth rate, some price is too high as the growth rate cannot stay forever (or at least the chance of that is pretty slim).

Randy,

Great framing. I personally said:

Unlike Saul, I do actually look at PS ratio for every company I own. I look at it like a price tag. SHOP goes for 19, Wix only costs me 9. Those are the prices. But that’s all they are.
…
You have to be careful not to think that a lower PS means a “better bargain.” Lower PS might just mean less impressive company.

http://discussion.fool.com/ps-ratio-specifically-33046009.aspx

That’s the way I look at PS ratio or “price.” But I believe Saul when he says he doesn’t look at PS. I just think he must have a good gut feel for it. If he were to look at a company like Shopify with $700 Millionish in revenue, and the Market Cap was $50 billion dollars (share price of $500 or so), I think alarm bells would start going off in his head.

Sorry to sound like I’m trying to speak for Saul, but I just think it may be difficult for him to quantify how he looks at valuation, because really it’s simply that his gut is well-atuned. The overarching theme of Saul’s investing is not finding the best bargains, but rather, paying up for the best companies, almost regardless of valuation. But in doing this for 30 years, I’d bet good money that he’s learned to sense when valuation becomes stretched too far (e.g. dotcom bubble).

The takeaway for me is this: There’s no formula to tell you when a valuation is stretched too far. The only things to do are:

  1. employ common sense (and hopefully a good gut), and
  2. don’t let any one position get too big, no matter how much you love it.

But that’s just my gut, so I could be way off.

Bear

22 Likes
20

I figure I might as well add some thoughts on this topic. FWIW, I’ve been in Enterprise SaaS sales for over five years, working with three of the top 10 SaaS companies.

Compensation:

  1. Reps are paid, in the most likely scenario, based on the MRR (monthly recurring revenue, my firm switched to this model last year). MRR is the monthly revenue that is recognized from the value of the subscription deal. If you have a three year (and the majority of all SaaS contracts are three year deals) with Total Subscription of $432,000. One year is $144k, the MRR is $12k. The rep is paid on the $12k. In traditional times, we were paid on the $432k. However, the payout is roughly the same (higher payout percentages with low dollars make you whole).
  2. The contractual terms of the software agreement have clauses that it will automatically renew unless written notice is given in “x” days before contract expiration. Typically these renewals are handled by a renewal team or an account manager, not a sales person (who is going to be paid to find the next net new deal). If you choose to terminate in after year 1, you will still pay for years 2-3. SaaS companies have contractual language to prevent this. The reason being is that SaaS companies, on the subscription only, are not making profit in the first three years. The average payback time is roughly 4-5 years depending upon the model (and why many of the SaaS companies are not profitable their first couple of years).
  3. Implementation services, which Tamhas mentions, can be anywhere from 6 - 24 months. Where I am, average implemenation is 12+ months with contact value about 1.5x-3x the software subscription. If you work for Oracle or similiar firms, implementation services don’t matter since they only sell software and rely on third party folks for services. Where I am now, we get paid a flat rate on the total value of the implementation services, day 1, and not over the course of the implementation. This has been true in my various roles in the last number of years.
  4. So if you are modeling something, the subscription starts on day 1 of contract signature. It doesn’t start when they go live or any later period of time. Implementation services are recognized the month the work has been incurred. For modeling, you can take the total implementation and divide equally by number of months - it won’t be 100% accurate but close enough.
  5. Retention Hypothesis - I would agree and disagree with you Saul. I agree that with high client satisfaction, folks won’t change. However, it is much, much easier to switch vendors upon contract expiration these days. Why - since everything is in the cloud, you don’t have any of the onpremise infrastructure costs, people you have hired, etc. to run the enterprise application. I was a SF beta test at a small firm years ago, we implemented SF, ran it for a year or so, then, because of political issues and a new client named Microsoft, we switched over to Microsoft CRM. Our transition was very easy, download the data and have someone put it up into the application (obviously work is done on learning new system, implementation, etc.) but my point is that the switching was relatively easy and a lot less painless than going from a legacy on prem model to a new subscription model.

I’ll also preface that this is for enterprise software sales, ERP specifically, and may or may not apply to some of the storage companies we have discussed.

Sox - who is hoping the Bruins show up for game 7 tonight.

20 Likes