Building Buffer’s Model

Buffer is a trailblazer when it comes to openness and transparency in the startup world, so they were the obvious choice for a great SaaS company to use as an example model using Opstarts scenario planning. Here’s how we were able to build a realistic model using that business data Buffer has generously shared with the tech and startup community.

We started with this very honest and open post from May, which went into detail on their current situation and plan. That gave us a starting May cash balance of $1.3M to use in the model.

Employee expense modeling: Buffer shares everything about their salaries. So to start the model, we imported their employee list from this spreadsheet. We didn’t find any data on payroll and other employee taxes, so we added a standard 15% employee overhead factor.

screen-shot-2016-10-19-at-3-08-11-pm

We also added a placeholder employee “Additional Employees” to account for employees that are no longer in the spreadsheet and use to model future hires. With ~80 current employees and a goal to double revenue in 12-18 months, we modeled new hiring of 3 employees per month in 2017 to support that growth.

Revenue modeling: Buffer shares both monthly reports and a revenue dashboard. We first imported their entire product list with subscriber quantities from May, and updated products with their actual selling prices (probably due to discounts or other custom deals, some of the product lines had revenue per user numbers that were a little different than the listed price). The vast majority of Buffer revenues are from six subscription products: Pro monthly/yearly, Small Business v1 monthly/yearly, and Small Business monthly/yearly, so we focused on just those products for this model. We looked at the May through September subscriber numbers for each of them and modeled some approximate growth and churn rates using those (in a real model, we would import in the exact subscriber data for a more accurate model). We modeled the Small Business v1 products churning with no new growth. With access to detailed customer data, we could instead create a link to model those users switching over to the v2 products over time:

screen-shot-2016-10-19-at-4-04-57-pm

But rather than guess at that, we kept things simple and just modeled those older product subscribers churning away. We modeled the v2 products growing very rapidly for the next few months as subscribers transition over, then stabilized the growth at 10% monthly growth over time. For the Pro products, we used 4% monthly growth rate for both products, with 2% of the monthly subscribers churning each month and 20% of the annual subs churning each year.

Expense modeling: We didn’t have access to recent data here, so we made a spreadsheet using numbers from Buffer’s transparent pricing page and imported those as the expenses. Since at the start of the model in May they were about double the revenue compared to that post, we doubled most of the initial expense values. For Compose and AWS, we modeled them increasing 2% each month. Healthcare costs we modeled as a link of 3% of salaries. Stripe we created links based on sales, 2.2% and 30 cents for each payment. Computer costs we modeled based on new hires, and internet costs we modeled based on number of employees.

screen-shot-2016-10-19-at-4-21-10-pm

 

We then compared the modeled numbers to the latest numbers published by Buffer and in their revenue dashboard. For September, they have $1.56M in the bank, and our rough model shows $1.61M. And our model shows $960K MRR in September vs $985K in their revenue dashboard – this is to be expected, since we only modeled growth in their main six products. But overall, this is pretty close to reality, and a great way to illustrate SaaS planning concepts.

screen-shot-2016-10-20-at-12-58-53-pm

Huge thanks again to the Buffer team for sharing all this info, it’s an amazing asset to everyone in the SaaS world.

Sign up for a free trial of Opstarts to try out scenario planning for your business.

Questions about this model, or want to learn more about modeling with Opstarts? Want to check out this model? Just email or tweet us and we’re happy to share!

 

Leave a reply