Buffer has helped startups by making everything they do transparent. This post analyzes Buffer’s public financials with the goal of helping you plan yours.
This post was originally published on 09 November 2015. Stay tuned for an update.
Recently, a social media software startup Buffer posted a finance planner job opening with the following questions (paraphrased), and this post is going to focus on the first one.
- Measure KPI’s for the business to ensure strategies are effective in driving the business.
- How can we better forecast our growth and understand what might happen 3, 6, 12 months down the line?
- How do we think about our budget more strategically? Where can we be more efficient?
- How do we align hiring objectives to overall business objectives?
- How do we think about new products and services and run cost/benefit analysis?
Now, to give a detailed analysis you’d need 1) full access to Buffer’s finances, traffic, and conversion data, and 2) work with them over time to start recognizing patterns.
We’re left with publicly available data, but we should still be able to make this post helpful, considering these are fairly typical questions for startups about to get their first finance hire.
Buffer’s Key SaaS Metrics
Andreessen Horowitz published a very helpful blog post outlining a common list of startup metrics, and most of these are particularly useful for SaaS. In this post, we’re going to focus on:
- MRR (Montly Recurring Revenue) and its growth
- Gross Profit
- CAC (Customer Acquisition Cost)
- Customer Churn
- LTV (Customer Lifetime Value)
Revenue is one of the key metrics to figure out if there’s a business in the first place. Specifically, is the revenue growing?
Buffer’s compounded Month-over-Month (MoM) MRR growth has been 4.9% over the last 12 months, amounting to 88% annual growth. Their Net Revenue — which typically includes revenue from professional services — has grown 78% in the past year.
But are these numbers any good?
David Skok and Pacific Crest Securities publish an annual SaaS Survey, which is an excellent resource for any SaaS startup looking to benchmark their financials. Here’s a graph from the 2015 survey illustrating revenue growth as a function of the size of the surveyed company.
Buffer’s revenue puts them into the $5MM-$7.5MM size range, with revenue growth rates ranging from 61% to 84% for the middle third group. With Buffer’s growth, they’re comfortably in the top 33%.
Note: MRR and GAAP revenue are not synonyms, but their revenue recognition is similar enough that we can compare them here. Even better benchmark would be data from Buffer’s industry, but such granular data for private companies is rarely available.
MRR Growth Per Plan
To further understand how Buffer makes money, we need to look at their finances on a plan level. It’s unlikely that every plan is created equal, and just a quick charting of the data can help us to find out where to focus.
As we can see, Buffer’s two biggest revenue-generating plans are Awesome at $10/m and Small Business at $50/m, and their annual counterparts $102/year and $510/year.
Here’s how each plan has grown over the last twelve months and their respective share of Buffer’s MRR.
While the Agency plan is growing the fastest, it contributes only 3% of the MRR, rendering it not significant enough to analyze further (yet). However, the Small Business Plan is growing almost twice as fast as their entry level Awesome plan and it contributes 1/3rd of Buffer’s MRR.
These findings make an interesting starting point to dive deeper into Buffer’s metrics.
Note: If we had full access to Buffer’s finances, one should compare their billings to deferred revenue to get a better picture of their financial health. For further reading, take a look at a primer on SaaS valuation by Scott Kupor and Preethi Kasireddy.
One of the most common startup financial mistakes is too strong focus on revenue while forgetting gross margins.
Gross margin matters more [than revenue] because it is directly tied to a company’s ability to spend to grow and achieve profitability. –Tomasz Tunguz
Just like with calculating key metrics, the first thing we need to do is figure out what belongs to SaaS businesses’ Cost of Sales. If you don’t know already , take a look at Jason Cohen’s detailed Quora answer. In short, CoS should include the cost of hosting and tech support.
With Buffer, we know their hosting costs in October 2014 from their transparent pricing blog post, and their customer service personnel cost from Buffer’s public salaries, updated on September 2015. We can approximate the current figures by calculating the hosting cost per customer from 10/2014 and multiplying it by the number of current customers.
We arrive at 80% Gross Margin, which is solid middle ground compared to the median gross margin of 78% by the 2015 Pacific Crest SaaS Survey. Ideally, we’d examine Gross Margin over time, but with only two datapoints with Buffer’s salaries, the only comparison we can make is to 10/2014 and 75% GM.
While already helpful, MRR and Gross Profit alone don’t give us enough information about the viability of Buffer’s business. What if they spend 5x the revenue a customer brings in to acquire that particular customer? 10x? For this, we need to analyze Buffer’s Customer Acquisition Cost (CAC).
Customer Acquisition Cost (CAC)
Unlike traditional software companies, SaaS companies incur customer acquisition costs upfront, whereas revenue is generated slowly over time. Ideally, we should calculate CAC for each paid acquisition channel, but with the public data available we need to get by with the “all-in” or Blended CAC.
The formula for Blended CAC is quite simple:
Where t equals the time period, say, September 2015.
Buffer does not have a sales force, and only one of their employees is a full time marketing person. However, since Buffer’s blog acts as a major customer acquisition channel, we should include the salaries of Buffer “Content Crafters” in calculating the numerator. Further, I’ve estimated a product team of 4 is responsible of their lead generation tool Pablo. Refunds, coupons, and onboarding expenses should be all included as well.
The resulting CAC seems to be suspiciously low, as the Gross Profit based payback time would be just over three months.
This could be due to the relatively early stage of the company, where their main customer acquisition channel hasn’t reached the saturation point. In addition, Buffer doesn’t employ sales or do a lot of paid advertising, which further points out that the CAC is only going to increase going forward.
Indeed, if we include Buffer’s four open marketing & business development positions (assuming all of them are filled tomorrow), the CAC would shoot up to $46 with a payback time of four months. This is closer to the median value of five months reported by the Pacific Crest SaaS Survey, for companies with Internet Sales as a primary acquisition channel.
It also looks like the team at Buffer estimates the saturation point for the blog might be approaching, as they recently announced a pivot on their blog content to focus on “Deep Tactical” blog content.
Going forward, I would recommend calculating CACs for each of the paid acquisition channels at Buffer. The Blended CAC is a good approximation and a starting point, but the relationship of putting down $46 and getting another customer is not as clear as with, say, Adwords.
At first glance, Buffer’s churn figures would appear alarming. They’re churning out 5.6% of their customers every month, or to put it differently, they seem to lose half of their entire customer base every year. In terms of revenue things are slightly better at 4.9% churn a month, but only slightly.
At a plan level, it looks like the Awesome Monthly takes the blame by having the largest user and revenue churn, followed by the Small Business Monthly. Their annual counterparts are performing significantly better — in fact, Small Business Yearly (SMY) experiences thecoveted negative churn. While 99.1% of the SMY customers are going to be retained in the next month, the remaining customers still make 100.9% of the original cohort’s revenue.
Negative churn is Awesome news for Buffer (no pun intended), and we’ll certainly get back to this later in the blog series.
The very next thing we should do is to run a cohort analysis for every plan, and see when these customers are leaving. Freemium users could have briefly switched to Awesome Plan, decided it’s not for them, and leave after a month or two. This would be significantly less troubling than their long-term customers leaving, in which case Buffer should reach out and find out why.
We would also want to know if the more recent cohorts perform better than the old ones (see the hypothetical graph above), as it would indicate Buffer is getting better in retaining new customers through improved onboarding or just generally a better product.
This would make a very interesting analysis, but unfortunately there’s not enough public data to perform this, so we leave it to the team at Buffer. Similarly, if we knew Buffer’s Net Promoter Score (NPS) for each customer, our analysis would get even better predictive abilities.
Nevertheless, the churn data by plan gives us enough ammunition to calculate customer lifetime values.
Customer Lifetime Value (CLTV or LTV)
Buffer publishes their customer lifetime values, but the vanilla version of the metric significantly overstates the value. This is fairly typical for the customers of out-of-the-box analytics services, since they merely pull data from their customers’ payment platforms without access to the cost structure.
Using revenue in the numerator overstates the generated cash flow from a given customer, while using just the churn in the denominator fails to consider sudden changes in the churn rate going forward.
But what should we do instead? There is a plethora of ways to calculate LTVs, and a16zsuggests using 12–24 months as a customer lifetime, whereas Jason Cohen recommends using the discount rate method. David Skok offers an even more advanced method of using churn curves taking into account negative churn. Here’s the formula with the discount rate:
If we don’t go the most advanced route, my personal favorite is the discount rate method due to its’ attempt to take into account time-value of money. However, estimating (guesstimating?) the rate can be tricky. I’ve used a 1% monthly discount, which amounts to 12.7% per year.
Next, we should provide context to the values, and the best way to do it is to compare them to CACs.
A good LTV:CAC ratio would be 3:1 or over, and depending on the method of your choice, Buffer’s values range from solid 3.3x to 4.4x.
Next, we would calculate LTVs per each of the Buffer’s plans, and compare them to their respective CACs. As we don’t have the plan level CACs, we’re left with LTVs only.
On the plan level, the differences in LTVs become even more prominent. In particular, Awesome plan with Discount Method has one third higher LTV than ARPU. While the relative difference decreases with the other plans due to their lower churn rates, the difference is still in hundreds of dollars.
With all this LTV soup, I can’t help but agree with Ron Gill, CFO at Netsuite when he describes in a recent a16z Podcast how there’s a real lack of standards in SaaS metrics. Hopefully this will change as the SaaS industry matures further.
Buffer’s solid revenue growth and high gross margins provide foundation for a healthy business. Their entry-level plan “Awesome” shows signs of early saturation, and it has a high churn, but the Small Business Plan is growing fast with lower churn and significantly higher lifetime value.
If I had all the data Buffer has, this is what I would look into next:
- Calculate CAC for each paid customer acquisition channel to see if any of them can be scaled profitably
- Calculate LTV to CAC ratio for each plan to find out where to focus efforts
- Start tracking customer retention per cohort to perform more advanced analysis
- Tie NPS scores to customer retention to improve predicability
- Start following all these metrics over time to find patterns and trends, and prioritize efforts into plans that show signs of product-market fit
Metrics analysis should not be done in isolation, and later in the series I’m going to look into forecasting growth and estimating Buffer’s burn rate.