What I Learned Selling A Software Business

Hiya guys!

Patrick (patio11) here. I apologize — it’s been a while. If you’re still interested in getting emails from me about making and selling software, great!

[Edit: Actually, it's possible that you've never gotten an email from me. Somebody might have just given you the link to this page, which is an online archive of an email that I sent to folks who had asked for it. If you'd like to get articles like this in your inbox, totally free, about once a week or two, give me your email address.]

I’ve been probably best known for running a teeny tiny little business called Bingo Card Creator. I sold it last year, and wanted to tell you a little of what I learned.

A Brief Background on BCC

It’s possible that some of you have never heard of this, so a quick story in-a-nutshell:

Bingo Card Creator was my first business. I started it in 2006 as something between a hobby and a dare. It makes bingo cards for elementary schoolteachers. Originally it was a Java Swing app. Since 2009 it’s been almost exclusively a SaaS app running on Rails.

The business model was always a one-time purchase, most recently for $29.95. Against all expectations, I got very good at organic SEO and AdWords optimizations for BCC. In good years it sold nearly $60k of software. In more recent years, rather less. BCC historically had a profit margin of around 60%.

I blogged extensively about running BCC from 2006 through 2014, with rather fewer posts since 2012 or so. It’s basically the root cause of anyone on the Internet knowing who I am.

Why Sell?

People often ask me why I sold Bingo Card Creator.

The biggest reason was focus: every year I found myself saying “Hmm, I totally neglected BCC this year, but next year I will try to reverse its slow decline.” And every year I found myself drawn into other projects instead. BCC was not taking up much of my time (thanks to Sugar handling most of the support) but it was taking more than zero of my mental cycles every year.

And what was that getting BCC customers? In round terms, well, nothing. I hadn’t made a significant change to BCC for years. Any CS request getting past Sugar (say, a request for a refund) took a non-deterministic amount of time — sometimes a day and sometimes a week or more. Content was not getting added to the website for years, due to a freelancer quitting and me not spending enough time to identify and onboard a replacement.

Starfighter was also coming down the pipe when I sold BCC. Thomas and Erin invited me to be CEO of a company that I was (and am) very, very excited about — fix dev hiring by writing video games! I’m so in! This meant for my other businesses: I’m so out!

Finally, like disposing of any asset, money was a consideration. Selling BCC was going to pay for living expenses while we built Starfighter’s first game (Stockfighter) and also pay for some development work to assist with the sale of my other SaaS business, Appointment Reminder.

These issues seem to be broadly representative of why my friends sell companies. Most of us in the fraternity of small SaaS shops sell when we have a better project to work on and/or when we’re just done with a particular product.

Why not keep BCC running in maintenance mode? A totally defensible choice! BCC was generating “a meaningful amount of money” (low tens of thousands in profits a year) without requiring a heck of a lot of work for it. That’s not an annuity, though: there is the chance in any given year that Google would change its mind about my site, that hackers would exploit it and ruin me, that a comet would hit Rackspace’s facility and require me to drop what I was doing and rebuild it, etc etc.

Profits are a risk premium, almost always. I’ve spent my risk budget and beyond on Starfighter, and didn’t need to have additional risk from other SaaS businesses.

Why not shut it down? BCC was my baby before my actual baby was my baby. I would have been enormously sad to see BCC disappear from the Internet and/or constructively disappear (like, e.g., turning off support for it).

What’s The Going Rate For SaaS Businesses?

A billion dollars.

Oh, whoops, BCC is not a Silicon Valley unicorn growing revenues at 400% for each of the last 5 years. Ahem. OK then: 2~4X TTM SDC.

TTM is “trailing twelve months.”

SDC is “seller discretionary cashflow.” Basically: revenue minus expenses which are both a) required to run the business and b) not paid to the owner. So, for example, the owner’s salary isn’t counted against SDC, interest expense is not ordinarily counted against SDC (not required to run the business since a hypothetical frictionless spherical SaaS owner would have no debt), travel is not counted against SDC except to the extent it’s required for the business, etc.

There is a lot of daylight between 2X and 4X. In general, this is where the intangibles of the business come in. People buying SaaS companies prefer:

Established businesses: Businesses get progressively de-risked with age. A company established 6 months ago might be doing something hinky that Google hasn’t figured out yet. They may be riding a novelty wave. They may be getting early adopting customers who are not representative of the total market. They may be accumulating technical debt at an unsustainable rate. A company established for 6 years has demonstrated that it is well past most of these risks.

BCC had 8+ years of history when it was sold, so this factor broke in my favor.

Recurring revenue: The best kind of revenue! People buying small Internet businesses are largely looking to purchase reliable cash machines. Nothing beats a SaaS business with monthly recurring billing on the reliability of next month’s revenue: given that it was $20k last month, it will be $19k+ next month with a certainty approaching that of next year’s NBA championships being won by a basketball team.

BCC had historically had a one-off purchase model, so every month we had to claw our revenue up from $0 all over again. This factor counted against me.

Note: Annual prepays are well-loved by SaaS owners but not loved by SaaS purchasers. Why? Some folks, when contemplating a purchase, actively promote annual prepays to (essentially) stripmine the next year of revenue from the business, then sell, which tends to a) inflate the TTM revenue (when reviewed by a less-savvy buyer or broker) and inflate the sales price while b) decreasing the MRR going forward and c) sticking the new owner with the cost-of-goods associated with the unearned revenue.

This is frowned upon. If you’re selling a SaaS business in the near future, don’t try to juice annual subscriptions beforehand — it will not end well for you.

Low technical risk: Surprisingly to me, the majority of people who buy SaaS businesses are not themselves technical. (This is less surprising if you think of SaaS businesses as SaaS businesses rather than SaaS businesses. Most people who buy McDonalds do not themselves have a huge amount of experience flipping burgers, right?)

This counsels buying businesses with well-understood technology which will require updating infrequently and not break in unanticipated fashions.

Practically speaking, this means that the market wants your website to be on WordPress and your technology stack to be something very, very mainstream, like Rails or PHP. That ensures that, when they want to hire someone to work with it, that person will exist in quantity and be inexpensive. The market highly dislikes software which is written in something more esoteric (the new Javascript hotness, Golang, Erlang, etc) because that adds substantial risk and cost to operating the business.

For similar reasons, buyers do not want to buy something which is hugely technically ambitious. CRUD app: great! An engineering feat of genius: problematic! Unless you have the engineers on staff and ready to transfer with the company, that’s selling the new owners a time bomb.

BCC had a very mature code base and mostly sane choices made (a Rails app… a creaky, ugly Rails app, but still a Rails app), so this broke in my favor.

Low market risk: Buyers want an asset which will continue being valuable in the future. This means applications which have constant or growing demand for them.

One can easily imagine software apps with declining demand for them. Games are the paradigmatic example: typically, they have a burst of sales at launch and then they depreciate straight off a cliff. Add-ons are great software businesses if the base product is something like WordPress, which has gone from strength to strength. They’re not very attractive if the add-on is e.g. for MoveableType, which is far less popular these days and getting less popular with each passing month.

BCC was pretty stable in terms of demand (“Do teachers in the United States still teach reading?”), so this broke in my favor.

Not a lemon: Back in the day someone won a Nobel Prize for pointing out that, if a population of goods has unknown potentially costly problems, and there is no way to determine which particular instances of the goods have the problesms, the market will penalize all goods in that population. The canonical example is used cars.

Small Internet businesses are, unfortunately, a market for lemons. By default, they're poorly documented pools of unknown, highly concentrated risks. The seller has an incredible informational advantage vis-a-vis the buyer about material facts like e.g. is the business presently penalized by Google or likely to become so in the future. Often, Internet businesses are not run in a very businesslike fashion -- expenses are paid in a haphazard fashion and not adequately documented, the owner is actually key to the business (for example, for technical firefighting) but does not disclose this fact, paperwork regarding IP ownership is not in order, etc.

All of these issues counsel acquirers of Internet businesses to assume they're taking on more risk than is being represented to them and decrease bids accordingly.

BCC is not a lemon: while there exist substantial risks in the business, they're largely the ones that were disclosed rather than ones waiting in the wings. Few businesses have a million words of backstory written about them contemporaneously. It was highly unlikely that I would attempt to cheat the buyer for advantage, as I have a professional reputation which is worth far more to me than N% of the transaction price. BCC had reasonably accurate books and copious documentation of well-understood business metrics (conversion rates, etc) dating for years.

How Does One Actually Sell?

Use a broker. Use a broker. Use a broker.

If you don’t have someone lined up already to buy the business, you probably want to sell through a broker. The primary value add is them pitching the business to a stable of customers developed without your involvement. They act as a firewall between the flakes ^H^H^H^H^H^H prospective customers and you.

Brokers will also help you in putting together a prospectus (a brochure with copious detail about your business, designed to make it look attractive to potential buyers) and help you with some of the mechanics like escrow and delivery. What you’re really paying for, though, is identifying the customer among a sea of flakes.

Have I mentioned the flakes yet? Oh, God, the flakes. So many flakes.

You know how, if you publicly post a job listing for a software engineering position, you will virtually invariably have 90% of the resumes be from people who cannot program at all? Imagine how flaky the pool would be if there was both an expertise requirement and they needed to be ready to write a $X0,000 check and they needed to follow-through on a complicated commercial transaction in N weeks.

People try to buy software businesses with “no money down.” (“Will you loan me the entire purchase price of the business? I’ll pay you back over the next 3 years. Promise!”)

People try to buy software businesses without exposing their own name, “for privacy reasons.” (I will get hate for this observation on Hacker News from someone who, despite possibly chronologically being an adult, is deeply skeptical of responsible adult businessman rituals like “signing contracts” and “wiring money.” Rock on, you beautiful snowflake, but no sane owner of a SaaS business will sell it to you anonymously. No, no, just throwing a blockchain into this transaction doesn’t substitute for several hundred years of practice doing this the usual way!)

People try to buy software businesses despite insufficient technical ability to operate Gmail.

People try to buy software businesses and get through weeks of negotiation about them before asking for a 40% discount literally when the final contract is in front of them. (This literally happened to me. Luckily for me, my broker had already identified a backup buyer.)

I used FEI to sell Bingo Card Creator. A+++ with regards to that transaction; would do business again. (If this suggests something about what mailing list you should be on with regards to opportunities about other software businesses, well, cough.)

Broker fees depend on the size of the business, generally in the 10~15% region. BCC’s was about 15%. This was entirely justified given the degree to which they de-risked the transaction and got a better multiple than I would have expected if I were handling the negotiation.

What Does the Timeline Look Like?

BCC sold very quickly, in only about five months from start to finish.

I contacted FEI to get the ball rolling in January 2015, and told them I’d be ready to work with them in earnest in February. January I mostly burned on a bunch of maintenance tasks to decouple BCC from the rest of my businesses.

Pro-tip: Do not re-use bank accounts, business books, or SaaS accounts between two businesses if you want to sell them independently of each other. This required massive work to reverse, including e.g. rebuilding the BCC deployment on Digital Ocean because a certain cloud provider fanatically supports moving images between accounts by not supporting moving images between accounts. I also had to do everything from migrate all of my email in Google Apps for Work (oh God, don’t ever do this), delete emails that the new buyer shouldn’t have access to, split various analytics and email providers apart, etc etc.

After about a month of preliminaries, I spent most of February working on the prospectus with FEI (~5 hours over a week; several hundred questions) and about a solid 40 hours on due diligence.

Due Diligence

Due diligence combines the length of a proctological examination, the fun of watching every season of the Wire, and the exactitude of doing taxes.

No wait, I got that mixed up. Due diligence combines the exactitude of a proctological examination, the length of watching every season of the Wire, and the fun of doing taxes.

It’s not enough to just assert “We sold $X0k last year and had profits of $Y0k.” You have to give primary documents supporting all of this, which will (in many cases) be “screenshots of websites” (Paypal, MailChimp, etc). Not just one screenshot, either: one per report you need to generate, which is probably one per month of the TTM. Times the number of relevant accounts you have.

Now do the same with your books. Paid $10 for SaaS? Describe it and why you paid for it.

Now here’s the really annoying part. This will all be checked. And even if you’re not trying to hide anything, you will be evaluated on the assumption you are. Which is how I ended up explaining to an ex-Goldman Sachs employee a) exactly how BCC accounts for AdWords spend (with Excel model to prove it lined up with Google’s calculations) and b) why a single chargeback in July 2014 was not recorded in the revenue books (“Simple error.”)

I’m told, against my expectations, that BCC was impressively well-documented by the standards of other businesses its size. This implies that many people are running their small projects in even more of a cowboy fashion than I do, for example by not having dedicated books for the business. If this describes you, God help you. At a minimum, get your books for the last year done professionally — whatever you spend on bookkeepers/accountants will be a pittance next to the time saved and additional valuation captured.

If you have good books and there is an expense in them which might questionably count against SDC or not, you’ll tend to win the argument, as good books solve most accounting arguments. If you’ve got bad books and due diligence finds a shortfall of $1,200 in a month, the presumption is going to be “Unrecorded expense; ding them for it.” Poof, there went ~$4,000 of the sales price. If you do a member contribution (cash injection from yourself into the business) of $1,000 and a member distribution (withdrawing cash) of $1,000, you’re highly likely to get the contribution discounted (because it’s not revenue) and have the distribution counted as a business expense (which it is not, but which you’ll have no way of demonstrating).


Hey FEI, close your ears for a second, OK?

Every broker of anything — houses, SaaS businesses, candidate introductions to engineering firms, etc — has a principal/agent problem. They’re compensated on a percent of the transaction, but it’s almost always preferable for them to get a transaction done right the heck now then to get the best transaction possibly done in the undefined future.

You might naively assume that brokers want to get the best possible price for the transaction and capture their percentage. The incentive runs differently: price for volume.

Brokers also discourage negotiation over terms because see above. They want as close to sign-on-the-dotted-line as possible.

In practice, buyers of websites might want some things which aren’t sign-on-the-dotted-line included in the “standard” deal, like:

Non-compete agreements: This varies heavily between deals. I was very, very happy to promise to not do another bingo or elementary school teaching website in N years. (“Want N to be 500?”)

Linking agreements: BCC owes some percentage of its link juice to its association with me personally. I committed to not removing existing links (like, say, the one on my kalzumeus.com front page) for a certain period of time.

Non-disclosure: The market term is “buyer gets to choose when/if to disclose the fact of the sale and the price.” I negotiated something similar to “I can mention the sale almost immediately after it happens as it will leak regardless; I agree to not disclose the price; you accept that BCC is extraordinarily well-covered on the Internets and that guessing the price will not be that difficult.”

The back-up buyer and I had our arrangement hammered out in under two days, starting from FEI's standard asset purchase agreement. I would ordinarily suggest getting the APA reviewed by your lawyer but I felt that it wouldn't have been cost effective.


You basically make a list of accounts that the business controls — Stripe, email, hosting provider, etc etc — and put them in a PDF file.

The business is then put into escrow, often accompanied with escrowing the domain names (via transferring them to the escrow provider). In BCC’s case, escrow happened but escrowing the names did not. (Rather than transferring domains to the new owner I put them all in a new DNSimple account and transferred control of the account.) Escrow makes the buyer post the sales price with a trusted third party, then instructs the seller to transfer the property (information/domain names). The buyer then gets to inspect the property and either approve the sale (in which case payment gets released) or dispute the property being as described (in which case mediators get involved).

BCC’s escrow process went very smoothly — it was all wrapped inside of a week.

Most sales of businesses will include a handover period — you don’t just dump a PDF on the buyer. You also give them a configurable amount of hours over a configurable interval to assist with transitioning customers, handle CS during their spin-up period, introduce employees/vendors/etc to the buyer, handle the inevitable post-transfer hiccups, and perform utility cleanup on stuff their technical staff might require.

I promised the buyer about 20 hours of my time over a month, and (for reasons mutually agreed) she ended up getting a bit more over about two months. The majority of it was spent doing things like going over marketing/SEO/etc strategy with her, making some not-too-difficult changes to the site, and helping her select and onboard a freelancer to deal with ongoing maintenance and development.

How Does It Feel?

git commit -m “Scrubbing publicly available evidence on the emails and about page that I was the founder of the company”

That commit was actually not entirely accurate (I didn’t rewrite history so much as greatly abbreviate it). It was not entirely accurate because the weight of what I had done hit me, all at once, when I was writing that. I cried actual, physical tears.

I think every founder has their own relationship with their business. My history with BCC was pretty intense. It was one of the only things keeping me going in my dark days as a salaryman. It got me out of that situation. It was a cornerstone of my public image for years and, sometimes, even my identity. I was The Bingo Guy.

Not being The Bingo Guy? That adjustment took quite a while but had mostly already happened. Admitting it? Oof, that hit like a ton of bricks.

Was it the right decision? Oh heck yes.

How Did The Transaction Affect Other People?

Employee: I employed Sugar (who handles the CS for BCC) through Pepper Virtual Assistants. It is unlikely in the extreme that BCC would have continued to operate without her services over the last four years. Accordingly, I decided to retroactively cut her in for 5% of the business. Props to Pepper for accommodating this request, as it is somewhat non-standard. ("Can you invoice me a substantial amount of money and promise me that you will pay a particular employee of yours a bonus of the same amount, net only of taxes?" "We can do that.") Sugar's participation makes the business much more salable than it would be otherwise, since she's capable of continuing to operate the business under new management (which decreases risks). I assisted the new owners and Pepper in transitioning her employment to the new firm.

Customers: I was adding no customer-perceptible value to BCC for the last several years of owning it, so I expect this is an unalloyed win for customers. They continue receiving everything they paid for and now have energized, responsive management at the helm.

Acquirer: BCC was purchased by Two Teachers, LLC. The principals are a married couple who are both long-time schoolteachers. They wanted to run a business (for freedom/lifestyle reasons) and were considering purchasing e.g. a Subway franchise, but a friend with an Amazon FBA business told them "The Internet is a much better option these days." (I don't know his reasoning but, if you asked me, I'd say "Vastly lower costs to get started; much better risk profile; more schedulable around other work/obligations than having 'business hours'; did I mention you don't have to cut sandwiches open for minimum wage because that seems important.)

Beth (my primary point of contact) and her husband plan to use BCC to learn the ropes of running an online business, and eventually have a small portfolio of products. This sounded like an excellent plan to me (BCC taught me much of what I know about business; the notion of recycling it to teach someone else was wonderful) and allayed one of my concerns about the sale, which was that my labor of sometimes-love would get strip-mined by a casino affiliate or similar.

How Is BCC Doing Under New Management?

I haven’t the foggiest clue. I sold the business so that I could stop thinking about it; I’ve stopped thinking about it. (I was about this close to putting it in my hosts file to prevent me from looking at it ever again, but figured that was cheating. I’m happy to report that I haven’t seen the site since the sale was finalized.)

I wish Beth and my former customers years of success and happy teaching, but this chapter of my life is firmly closed now.

To new adventures!


Patrick McKenzie

P.S. Now that Starfighter is out of OMG-we-have-to-launch-eventually crunch mode, I’m writing again! (I hope?) As always, drop me a line if there is anything you’d like to hear about.