Back of the Envelope: How to Estimate the Annual Revenues of Any Private Company | Hacker News Reader

Back of the Envelope: How to Estimate the Annual Revenues of Any Private Company

ryanborn.net

52 pointsaudiomicro26 comments

A dead simple way to estimate the annual revenue of any company.

Such a formula has so many caveats that it is not terribly useful in revenue estimation, except perhaps to provide a range within an order of magnitude. For example, Starbucks has $60,000 in revenue per employee, while DirecTV has $1.6mm. Surely no formula that would make predictions for household brands that would make errors anywhere from +66% to -94% is useful.

The main reasons for the variance is (1) how much value add the company generates for its customers, and (2) what portion of that value add is labor.

Some companies' products are very labor intensive, but they outsource most of that labor, so the company's product has high revenues per employee because there are "invisible" employees not on the payroll due to outsourcing arrangements. This is where the company is providing low value-add. For example, consider a trucking broker, who provides transportation services to its customers, but outsources the actual trucking to owner-operators who are not employees.

Other companies have high value-add, but very little of it is labor. For example, a television cable company has high costs for licensing media content, which is a big part of the value it provides for consumers, but that licensing content requires low levels of labor (unless they vertically integrate).

Companies that have high value-add, mostly in labor, tend to have moderate levels of revenue per employee. A consulting company is an example of this. It requires a lot of labor, but almost the entirety of the value add is that labor.

it's an ESTIMATE! simply adjust upwards for industries where the average salary is over $40k. How about this....ANNUAL REVENUE = NUMBER OF EMPLOYEES X AVERAGE SALARY OF EMPLOYEES X 2.5 . Try it now.

This is complete rubbish and shouldn't have made HN. Is it even for real?

I mean, yes, IF your employees cost 100k/year on average, then you need (100k * number_employees) to break even. (Thats what 'average' means!)

So IF you break even, your revenue must be at least that. (thats what 'break even' means.)

But theres no information here! That's not a bound, or estimate of your revenue, its just the amount you need to pay your staff, if your staff have average salary.

This 'formula' makes no allowance for some companies having a higher average salary than others. It makes no allowance for some companies having a higher profit margin than others.

Nothing to see here; move along.

it's an ESTIMATE! simply adjust upwards for industries where the average salary is over $40k. How about this....ANNUAL REVENUE = NUMBER OF EMPLOYEES X AVERAGE SALARY OF EMPLOYEES X 2.5 . Try it now.

This formula is almost useless. It gives an estimate 50x too low for craigslist, 10x too low for google, and even 2x too low for wal-mart. What use is a metric that's no better than a wild assed guess?

37signals has a related post:

http://37signals.com/svn/posts/2283-ranking-tech-companies-b...

I have not checked their figures but in the case of these companies, the proposed 100K formula is mostly way off the mark.

As for NASDAQ-100 companies, the following chart is very informative:

http://www.jbryanscott.com/2009/02/07/nasdaq-100-revenue-per...

I think the 37Signals post was meant to show outliers (how high it could go if you are awesome).

But that NASDAQ-100 post is insightful. It would probably be safe to say top companies exceed $100k per employee when doing well.

I've seen this equation used to estimate costs for private tech companies, but never for revenues.

And i have seen this equation used as heuristic to estmate the number of employees of company, when to hire and when to reduce. This is gross of course because the calculation should be done on the benefits. The production costs might be high and 100k per employee not enough.

There is no data to back this up at all. You can't take a set number like $100k and apply to all companies across all industries.

This wouldn't work for companies that have significant costs other than employees. So for tech companies it's generally accurate, but its unlikely to work for other industries.

No doubt!

Though I disagree that it would work for tech companies, since that encompasses such a broad range of companies.

A tech company that serves mostly enterprises may employ far fewer, yet more highly paid employees for customer service, yet a tech company that sells directly to consumers (especially if it's sales that requires a large customer service department) may employ many, lower paid employees.

The other problem I have with all of this is "why revenues"? I mean, that's a valuable thing to know, but without margin or profit, it's an easy figure to game. Ask kozmo.com or MCI|WorldCom or any of the "Generate lots of revenue then IPO" companies of the late 90's early 00's. Investors lost their minds and forgot that if you sell a loaf of bread for US$1.00 and you pay $1.50 for it, you've lost $0.50. You'd be better off not buying the loaf of bread (or, perhaps, getting out of the bread loaf selling business). It tells you nothing you can really trust.

(Edit: to fix basic math error)

I have used a similar formula for a long time. 30K x # of employees x 2 to get the gross profit needed to operate. Adjust revenues to match the gross profit of the industry. The author states 50% profit for companies, but some are much lower and some are higher. I have been self-employed for 20 years and most of my friends own small businesses (non-tech). This formula has worked roughly for all of them.

So... 60k x number of employees?

I'd like to see this tested on known datapoints--that is, public companies that disclose number of employees and profit revenue.

If someone has a bloomberg terminal handy it should be fairly easy to setup a worksheet to test it (both employee numbers and revenue are available as standard fields).

Am I the only one that dislikes the business model of the bloomberg terminal?

This is front loaded with a TON of assumptions. And you could easily make the argument that it doesn't apply to most web companies. Take craigslist for example...estimated rev of 122 million and around 30 employees.

If you insist on using this calculation for anything, best to stick to local electrician shops.

I remember a few years back it was around 230k average per employee of a public company (according to Forbes).

That number must be higher now.

As others noted, this number can range greatly depending on industry.

yep - you must adjust based on the average employee salary based on the industry

As with any back of envelope calculation it makes up for its lack of accuracy with speed and ease of use.

It just gives you someplace to start an analysis. It's not an end in itself.

There's nothing preventing the application of additional knowledge about a specific industry or company to modify the basic equation.

The last company I worked for paid me market rate as a Rails dev, but they made almost zero money. They had a monthly burn of £50,000. In other words burning through VC money. This equation is not a magic bullet.

It works like a charm...remember - it's a quick and dirty / back of the envelope calc. don't troll on it. It's an ESTIMATE!

Hm, this doesn't work for at least the company I work for. Without disclosing too much, he's off by more than half.

That is, in your case the formula gives the right order of magnitude. Great precision for such a rough estimation.

"For every problem, there is a solution that is simple, elegant, and wrong."

Reply on news.ycombinator.com