Analysing software businesses

I have analysed most of the software deals above $20m EBITDA in Europe over the last couple of years and have attempted to build, albeit far from complete, a framework of how to look at these businesses. I have spent over 10 hours with:  

  • Former CFO, Visma
  • Former, Account Director at Access Group
  • Current Head of EU at Automation Anywhere
  • Former Commercial Director at Open GI

And here is what I have to show for it:

What is software and why should I care?
Software companies are those engaged in the development, maintenance and publication of software as a core business. Think Microsoft, SAP, or Oracle. Understanding software is not only helpful for analyzing large technology companies, but the structural design of the business sheds insight into the why companies such as AMZN, FB or GOOG are so profitable.

Also, if you want to work in private equity or late stage venture; it’s pretty much standard knowledge.  

The Private Equity wave in Software

The chart below from this BCG report says it all. The number of annual software acquisitions more than doubled, from 228 in 2007 to 481 in 2016.

Private equity deals are flowing. KKR sold Visma for $1.8bn in one of the biggest European deals last year. Montagu recently took healthcare ERP vendor Servelec private and are likely to follow the playbook of Vista followed with Advanced Software. KKR, EQT and CVC also all recently launched growth equity funds focusing on technology and are increasingly competing at late stage venture for smaller companies. So not only is a conceptual understanding of software crucial to grasping how and why technology businesses are the most profitable in the world, but if you want to work in PE you have to be able to analyse the 200 VC backed software companies that BCG believe are coming to market through 2021.

So what is the strategic rationale of buying and building software companies? How do the economics work of typical software businesses? Why is this space so attractive? This post aims to provide you with a framework of how to analyse these businesses.

Business Models

There are generally two software business models;

  • Perpetual license – customer owns that version of the software and then renews every 3-5 years. Software companies often sell a perpetual license with the first year of maintenance included in the price and the customer then pays ~20% of license cost as maintenance fee annually. Microsoft Office is a prime example.
  • Software as a Service / SaaS the vendor hosts the software in a cloud and sells the service as a subscription. Think Netflix or Google Apps.

Business models are increasingly moving to SaaS business models because it benefits the customer. Even though the total cost of ownership of the software between the two is similar, the cash flow profile for the customer is different. SaaS shifts laying out cash for a license (capex) to an ongoing pay-as-you-go model (opex).

Investors also prefer SaaS models for two main reasons:

  1. Higher predictability of future cash flow – SaaS has higher recurring revenue than license model. This provides a more consistent stream of cash flow with less ‘renewal’ risk at the end of every license.
  2. Cost structure – the larger the upfront license cost, the larger the sales team required. SaaS models usually have a lower sales and distribution expense than license models.

Without going into too much detail, software is employed either on-premise (installed and run locally on customer hardware) or in the cloud (info is stored in a data centre off-premise). The Access Group, a UK based mid market ERP vendor currently on the market, lays out when a customer should choose each deployment model here.

Operating Cost Structure 

There are a few major components of the COGS for software companies:

  • Hosting and maintenance costs (salaries, support centre etc)
  • Depreciation of capex if company has servers
  • Implementation and training costs
  • Sales and Marketing spend can also be bunched in COGS as in the case of Visma.

The gross margin of recurring revenue for software is normally in the region of 70-90% and increases with scale. Why?

Fixed vs marginal costs

Software is so profitable because it has large fixed costs and near-zero marginal costs. Understanding this concept will shed insight into the profitability of the powerful tech companies. Once Microsoft develop Office, they can distribute it to everyone. Once Facebook built the network, it is open to all advertisers. Likewise with SaaS players, once it has developed the ERP solution, the marginal cost of distributing is limited to sales and distribution of the product. This is operational leverage; total costs do not grow as fast as revenue. How you sell and distribute then becomes one of the most important structural elements of the business.

Sales and distribution

The sales and marketing (S&M) spend is normally significant and depends on the distribution strategy. There are direct and indirect channels. Direct is driven by internal sales teams and is normally to serve large enterprises or public sector customers. The software vendor will pay a commission to the salesperson for license renewals or a certain percentage of first year subscription revenue. Indirect channels are normally to serve smaller, fragmented markets with many customers. Indirect can be in the form of partners who receive 25-30% commission.

When analysing various distribution methods, investors commonly ask:

  • What is the incentive to drive new sales vs renewals? How does the compensation structure differ between product lines?
  • Does the partner supply other vendors? What is the incentive of the partner in selling your product?
  • Does a high renewal commission imply it is hard to renew licenses?
  • Are partner commissions gross sales which have extra expenses? Does the third party implement the solution too? Are these commissions accounted for as net or gross sales?

After spending countless hours trying to dive into meaningless features of competing software products, I eventually realised that a large part of the competitive advantage of the most successful software businesses is normally around the sales network and process.

For example, Visma has a large partner network that serves over 12,000 SME’s in Norway. Partners receive 60% of first year sub revenue, 20% in year 2 which then decreases in time. Visma only charge the client twice in the first year so they actually have a negative cash flow on year 1 and only breakeven in year 5. But with an average lifetime of ~20 years for SME ERP in the Nordics, the run off-EBITDA margin can reach 35-40%. Around 90% of partners also only serve Visma which makes it difficult for competitors such as Microsoft to match the scale of Visma’s network.

It is also worth considering CAC when analysing new customers vs S&M in a period. See here for more on unit economics of customers. 


This is where software businesses achieve synergies in M&A. Main components are labour, rent, H&R, legal, etc. Visma aim to cut up to 30% in acquisitions within 1.5 years. Access Group aimed to double  EBITDA of the acquired business in two years by cutting expenses and cross selling. The G&A base doesn’t scale as revenue changes. Also, the financial flexibility of the cost base is important to consider when the company can face issues. For example, the capacity utilisation of the labour base. If customers drop by 10%, do you still need the same number of workers?


R&D is the real invested capital for software companies and normally needs to be adjusted when calculating ROIC. R&D is accounted in two different ways:

  • Capitalised – increases intangible assets and is normally  amortised over 7-10 years.
  • Expensed – deducted straight from revenue on the income statement.

Capitalizing R&D inflates EBITDA and thus when comparing companies it makes sense to deduct to normalise FCF.

Customer Retention

Retention for SaaS businesses, especially ERP solutions, is normally in the 80-90% region.

High retention = predictable cash flow.

Retention and customer churn are the most guarded metrics for subscription businesses. Averages are a dangerous game when it comes to churn.

For example, if 5 customers of 100 cancel in the past month, an average churn would be 5%. But if all 5 joined 2 months ago, there are issues with the product. Analysing by cohort gives a better picture of customer behaviour. Analysis by 1010data shows the difference in retention by month for the meal delivery companies:

There are many different reasons for customer stickiness; different levels of systems integration, mission critical software, cost to benefit ratio, and the time and cost of retraining employees. However, all senior executives seemed to think their vendor was stickiest and it wasn’t until I asked some detailed questions, such as below, that the vendors with sustainable retention became clear.

  • Usage and engagement metrics are important understand the quality of the recurring revenues and actual stickiness / value add for the customer.
  • Check the terms customers are renewing on. Has the vendor cut the costs of the service by 50% to maintain customers?
  • How does contract renewal work? (Pricing, sales commissions, process)
  • Higher customer concentration is inherently risky and low concentration is expensive to maintain
SaaS Unit Economics

Once you have customer stickiness, you can acquire new products and up / cross sell with the same salesforce. This increases margins at scale. However, it’s worth mentioning that not every $ of recurring revenue is created equally. For example, hardware and software renewals have completely different cost structures.

The unit economics of streams of recurring revenue gets at the heart of why PE love SaaS. The classic annuity model laid out below is a simple framework to calculate long run NPV. assuming an average EBIT margin of 65% on recurring revenue, we can analyse the implications of churn on SaaS businesses. If a company has 10% churn, this implies they will lose 10% of customers per year, hence the negative growth rate. The table below shows if the retention rate improves from 10% to 5%, the PV of recurring revenue from this customer base increases by ~30%. The amount of recurring revenue and the retention rate of this revenue is crucial to value subscription businesses.

Damodaran’s useful data set also shows how the unit economics translates into attractive metrics for the industry as a whole:

  • Return on Capital ~ 45%
  • Net working capital / sales ~ -11%
  • NOPAT margin ~ 25%
How PE creates value with Software

PE purchase software companies as a platform to ‘buy and build’. BCG’s sample saw bolt-on acquisitions increase from 47% in 2007 to 61% in 2016. Buy and Build is as it reads; PE funds acquire a software platform and look to bolt-on various other software products. HgCapital completed more than 120 bolt-on acquisitions with Visma from 2006-16. These bolt-ons ranged from Agda, a Swedish payroll software provider, to EasyCruit, a recruitment solutions provider, to Netvisor, a SaaS ERP vendor to Finnish SME’s. Hg both internationalised and embedded Visma further into each one of their enterprise customers.

Normally, a lot of the value create by PE is in internationalising and professionalising the sales and marketing of software companies. Especially the smaller mid market companies. The founders are normally engineers themselves who hire engineers to constantly refine the product rather than distribute it at scale. 

This buy and build strategy drives organic growth as enterprise customers prefer one rather than multiple vendors. By upselling to customers you extract more dollars per customer which drives organic growth. Visma saw revenue and EBITDA growth compound at 17% and 23% respectively over the 11 years that Hg has been invested.

Another reason SaaS businesses are popular with PE is because software economics match the return profile of of both VC and PE investors. Firstly, the original product with a fixed cost base plus increasing returns to scale earns a high ROIC and can scale with little capital. This matches the low-hit / high multiple return rate VC crave as they can pick the correct product and then sale with little marginal cost. PE then acquires from VC and provide the capital to acquire new products to bundle with the original offering. This strategy also matches the return profile of PE as they can acquire and add various products to the platform over the 5-7 average holding period of PE portfolio companies. Although the economics are not as good as VC stage due to the capital required, the risk is relatively lower as you have product-market fit and sticky customers.

So, in short, software roll ups are so attractive because:  

Operating leverage – high fixed vs variable cost base gives high returns to scale

Customer stickiness = predictable cash flow profile

Growth runway – mid-market / lower end of software market is very fragmented. Lots of different solutions to acquire that can be bundled into the solution and provides clear value creation plan.

Multiple arbitrage – average EV/EBITDA multiples are around 10 at entry and 13 at exit for PE funds. The average bolt-on acquisition for Access Group costs 6x EBITDA and Access is valued around 12x EBITDA. Visma followed a strategy of purchasing SaaS businesses for 3-4 recurring revenue and were acquired themselves at 5.7x total revenue. I guess it’s a pretty good deal. 

One thought on “Analysing software businesses”

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s