Hacker News new | comments | show | ask | jobs | submit login
Startup Metrics (a16z.com)
252 points by webmasterraj 787 days ago | hide | past | web | 48 comments | favorite

My colleague Ed made a cheat sheet for SaaS metrics [pdf] https://chartmogul.attach.io/NkqgtF8H

This is handy; thanks!

This makes me happy. Thanks!

Just curious, did something happen during Demo Day presentations that made both Sam and AH write a blog post about this problem?

I don't think so... I think a lot of people screw it up though.

> "How and when revenue is recognized is governed by GAAP."

Being a CPA & coder this statement really annoys me, GAAP for those that don't know simply means "Generally agreed Accounting principals" and there is really no acceptable reason for this acronymn to be used outside of the (accounting) profession. Maybe it wasn't the authors intent but it smacks of those people who use jargon to help aggrandize their position when a simple inclusive explanation would have been more appropriate.

You're talking about two different issues:

1) Whether or not people outside the accounting profession should use the acronym 'GAAP'.

2) Whether people use the term as jargon to help aggrandize their position.

I have no comment on point 2. On point 1, though, my position is the polar opposite of yours. Business people in general (and especially CEOs!) should understand the principles of accounting, the ways in which transactions are recorded etc. If they don't understand these things, how are they meant to understand their company's financial statements? If they can't understand their company's financial statements, how are they meant to manage the company?

BTW - I am also a qualified accountant, although this is not related to my current job. I feel so strongly about these issues, though, is that I periodically run an 'Intro to Accounting and Finance' course for interested colleagues, and have delivered similar training courses at other large tech companies in the past.

Actually I agree with your points, my comment was made in a rush and on reflection is full of hubris. My complaint was with the article which is attempting to define terms for newcomers not defining this term. I didn't mean non accountants shouldn't use the term, I meant that use of this term as an acronymn without definition in a definitions of things article seemed inappropriate.

GAAP is not an uncommon acronym in tech circles. For example, listen in to any earnings call or read any of the big tech companies' quarterly earnings press releases and you'll find references to GAAP all over the place. Quarterly earnings are a pretty common subject of discussion when they occur, so I wouldn't be surprised if a lot of programmers are familiar with the term.

I have to agree - I see it mentioned a lot throughout articles, tweets etc when discussing funding or earnings. I've never felt it was a particularly 'accountants only' term.

Correct me if I'm wrong here, but my understanding is that GAAP is not just stuff people generally agree on, but a highly technical set of rules:


And that they are determined in the US by the FASB, which is part of the SEC:


The SEC, of course, being the federal agency that regulates all attempts to take investment.

So I think the acronym's use here is important, because it signals that there is a precise legal definition of the term revenue, and that startup founders, so used to hustling and half-assing things, should not fuck around when using that word.

Sure, I really agree with you. My beef was not their use of term but the fact they didnt go on to define it in what is a let's define the things article. I suppose if they'd just unpacked the acronymn with a brief definition would have made this article more approachable in my opinion.

  GAAP for those that don't know simply means "Generally agreed
  Accounting principals" 
Are you sure you're an accountant? GAAP = Generally ACCEPTED Accounting PRINCIPLES

Yes. The term is not legislatively defined, it is a term used in industry, over the years I've heard it defined both ways as "agreed" and "accepted", but I agree that accepted seems more common now. Talks to the point of why use of undefined acronyms can lead to confusion.

As an engineer turned PM who might start a company some day, I've always seen financial literacy as one of my weaknesses. Companies are ultimately judged for their finances. Ignoring this subject is dangerous for a company's prospects, and illegal if it causes you to make misleading statements to investors.

If anybody is looking for further reading on this subject, I enjoyed "Financial Statements" earlier this year(http://www.amazon.com/Financial-Statements-Step---Step-Under...). Quick and straightforward, it starts with the vocabulary and explains the major financial statements (Income, Cash Flow, Balance). The majority of the book follows a company from inception to dividend disbursement, explaining how various business activities (e.g. signing a lease) hit the financial statements.

I would recommend a course in micro economics. Learning micro economics really changed how I think about finance. Anyone who plan to start a company some day should know how markets work, marginal return, opportunity cost, sunk cost, etc.

I really like the fact that this post first distinguishes "#1 Bookings vs. Revenue".

I hear it almost everyday among startups, where they talk about bookings, revenue and incoming payments interchangeably. And sometimes even in the same sentence such as "We made 100 k in the whole last year [meaning revenue], we currently make about 50 k per month [meaning bookings] and just last week we made another 30 k [meaning incoming payment]."

This makes it way to hard to properly communicate with founders. So great stuff this post.

Does anyone have an estimate on the average burn rate for a startup in San Francisco? Seems important these days, given talk of the end of easy money (see @bgurley)

This company's burn rate is over $400,000 now according to http://mattermark.com/how-we-spend-money-at-mattermark/.

It's abnormally risky to raise only enough money for 17 months of runway. I wish Mattermark the best, but this level of burn is not the most confidence inspiring.

From what we understand from investors, it is actually pretty typical to raise enough money for 18 months of runway. Additionally, we still had ~1.5m in the bank when we raised.

I will certainly report back on how it goes though!

I would think that the end of easy money would have a huge effect on the prices of things that startups buy.

In other words, if @bgurley is right and the economics of money-losing companies have changed significantly, then historical burn rate numbers (from 'easy money' times) are not going to be particularly useful to you.

Cutting burn rate is notoriously hard to do. Startups are built on dreams, and every single one of them has a longer list of things it wants to do than the capability to do them. Accepting that the status quo has changed means founders have to cut expectations of themselves and what they hoped to achieve in 6/12/18 months. Instead of cutting burn, many will start the hamster wheel of trying to raise money, even at unfavorable terms, and realize the reality only too late.

Short answer: burn rates will continue to be high until too late into the down cycle. So current levels matter, a lot.

I kinda assume that if easy money is over (and I don't know that it is. That would be my guess, but it's just a guess) - if easy money is over, then the startups that depend on easy money are probably mostly doomed, but that seems obvious and uninteresting.

I was talking about the next generation of startups... if this generation is, in fact, doomed, there are going to be, for example, more supply and less demand for small office sublets in the trendy parts of town

I believe one of the tricker question is whether your revenue is truly recurring.

E.g. you may say consulting services or just charge monthly support subscription.

Especially in short term it is easy to convince yourself that it is recurring (e.g. monthly access to tutorials) when really churn is so high, since people use it on-need basis rather than long-term.

Two questions:

How do you deal with refunds? Let's say a customer buys a product, for whatever reason their expectation not met, and a refund is issued. Do you adjust revenue in the month the revenue was recognized? Or do you add negative revenue to the month in which the revenue was received? Or something else?

And for average monthly growth rate, that can vary a lot for the same company depending on where you start. If a company has monthly revenues of 1, 6, 10, 12, 14, 15, 16, it has an average growth rate of 58% over the past 6 months, but 10% over the past three months (and 7% over the past month). I see startups using arbitrary start dates to inflate average growth rates, and am wondering if there's something standard when presenting this metric.

1.) Refund: Yes, I would adjust the revenue in the month it was recognized.

2.) Monthly growth: It is actually really simple, you just caculate:

(16/1)to the power of (1/6) which equals: 1.58 This means the average growth rate was 58 %.

16/1 is the total growth rate for all 6 months. And to the power of (1/6) because it is calculated among 6 months

1) For a business with consistent refunds in prior months (e.g. money back guarantee), this has the odd effect of always being able to show growth compared to previous months, even if revenue is constant.

2) Right, that's how I calculated 58%, but does it really make sense to calculate the base off of some arbitrarily low first-month revenue? Does it make sense that a company with strong first month sales should have a much lower growth rate than a company with abysmal first-month sales, given a certain current monthly revenue?

1) If you expect a certain refund quota, lets say 20 % within 6 months, then you should already account for it: only account 80 % of the actual revenue and adjust this number after 6 months when you now the actual refund quota. This makes it more complicated but it would be the right thing to do and will help you get a somewhat decent financial plan.

2.) If your growth curve flattens out than this approach obviously does not really give you a good representation of how well you are currently doing.

> Active Users

What is a typical or good (active users)/(total signups) rate for a SaaS company?

This is sensitive to your definition of "active user", and what service you're selling.

It's normal to see a lot of attrition right after users sign up, especially if the lowest tier of the service is free. More interesting is how many users you retain after the initial kicking of the tires.

Each product is different, even those that appear to be the same. I'd try and stay away from comparing yourself to others and instead look at the data the right way. Then I'd compare yourself to yourself each day and week and do things to improve the numbers.

Instead of looking at it this way, I'd instead look at it on a cohort basis over time. A product can have a really bad (total active users)/(total signups) rate historically but actually be in really good shape after iterating for a while. For example, you might be working on your product for 6 months, have 3,000 signups to date, and 300 active users. It would be unfair to measure how things are going by taking 300/3,000 (10%). If you were to look at the same data on a cohort basis, you might find that some group of recent users is stickier than others (because your product changed, or product positioning changed, or something else), and the rate might be something more like 90%.

Products like Mixpanel and Amplitude handle cohort analysis well. You can send in a couple events (Sign Up and some authentic usage metric you define depending on your goals) and pull a cohort analysis. An example: http://aacook.co/retention.png

This chart tells you quite a bit about how you're doing. Week/week acquisition (number of new users signing up) is in the first column, new user activation in the 2nd column (number of new signups who reached a moment of value) and a basic form of retention (number of users coming back at week N).

While great, this seems like not-so-easy work. Perhaps this is an opportunity for a company to charge a couple of hundred dollars monthly and create semi-automated reports for your investors.

Totally agreed. I've been consulting in this area for the past year now and it's a very common problem. One of the most powerful things you can do with these metrics is pull them on a regular basis and write updates. For example, write about how the week went and send it to the team. Or write about how the month went and send it to investors. Even if no one reads the updates, the exercise of just looking at the data and explaining yourself can really help you focus on what matters.

Tools like BareMetrics and ProfitWell do a nice job pulling revenue data via Stripe. Out of the box, tools like Mixpanel and Amplitude do a good job pulling giving you product and engagement metrics. It's up to you, though, to send all of the data over to those packages correctly and learn how to understand and interpret their reports. Other data, like burn rate, is often calculated back-of-the-napkin.

I'm working on a tool called Growth Report and shipped an MVP a little over a month ago. Hit me up if you want to check it out. aacook@aacook.co.

I do not have a product, and I have not raised any money, but if you have an MVP you should definitely post it here. This solves a real problem both VCs and entrepreneurs currently have. To the point where YC CEO and partners at A16Z (both leaders in their categories) have to write about it.

Usually this is best done by in-house engineers, I think. At least, that's what the founders at my startups did. You want to really be intimate with your metrics.

I suppose it could be productized, I am pretty surprised that there isn't a super-easy plug-and-play solution (a la Dropbox, Slack, etc). Probably because coming up with something that's customizable for different startups' needs must be really challenging.

They are called TempCFO, but cost more than that. Saved my bacon.

What sort of price range out of curiosity

unpopular amongst hackers, but really it takes a couple years of business school to learn real accounting, finance

Regarding ...

> #5 LTV (Life Time Value)

Shouldn't the calculation factor in the time value of the money over the estimated life time?


It's a pure summative estimate, as I understand it.

I imagine that it's taken as the input for a net present value calculation by many investors.

Two equal LTVs could results in pretty different NPV results depending on the expected lifetime, and using both averages for the discount definitely has some error.

That's a good point. I'm reminded of von Neumann's quip that "with four parameters I can fit an elephant, and with five I can make him wiggle his trunk."

Since there's so much accounting in these metrics, anybody has a recommendation for good accounting software, or knowledgebase? Especially for hardware startups (e.g. building physical stuff, inventory, keeping track of component and parts)...

great response to sama's blog post from earlier today

Thanks for that, brilliant. Sticked!

This is Entrepreneurship 101

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | DMCA | Apply to YC | Contact