« Beeminder home

Beeminder Blog

A dog gun shooting bees (it doesn't make any more sense than that if you view the actual image)

UPDATE 2021:
We’re pausing (most of) these dogfood bounties!

We still like them in theory but it turns out we all derail often enough that the logistics of getting money from us to users is more than we can keep up with! We do personally enter our personal credit cards in Beeminder and get charged money like anyone else and, yes, for those of us with equity stakes in Beeminder, the sting is diluted and that’s suboptimal. But! It’s surprisingly little sting dilution due to how much inefficiency there is. We lose a big fraction to taxes and other overhead and all of us — including us founders — have, individually, less than 50% equity stakes each. Meaning that when Beeminder charges us, say, $90, a pretty large majority of that $90 is truly disappearing for us. In fact, when we were paying those pledges as dogfood bounties to users it was kind of a warm fuzzy feeling (plus the marketing value, etc!) that diluted the sting similarly to how you superfans have the sting diluted by the warm fuzzy feeling of supporting Beeminder!

In conclusion, we believe in dogfooding as much as ever and we’re approximating it imperfectly but well enough by just using Beeminder exactly the way all other users do, even without these dogfood bounties.

PS: We floated this among Beeminder superfans and they nearly universally thought it sounded totally fine.

Excerpts from the debate with the dissenters

ZM: The bounties keep you honest! They have value emotionally and in terms of optics. This feels like one of those beautiful unique Beeminder scrupulous charms lost to the juggernaut of corporatization. Also there’s a huge difference between “this pledge goes to someone else” and “this pledge goes to my own company”.

DR: First of all, I don’t feel like honesty/dishonesty comes into play. We’re at least striving to be 100% transparent about it all.

As for that huge difference, it’s a fair point but note the surprisingly high fraction of a payment that gets lost to fees and taxes in the process of coming back to us as dividends. Even if we 100% identify with the company — like every dollar of benefit to the company brings us the same happiness as a dollar to ourselves — we lose a lot of money on derailments paid to Beeminder.

Finally, I don’t feel like there’s much of a slippery slope here or that we’re remotely selling out or anything. I mean, strictly speaking, we’re holding ourselves to the exact same standards, using Beeminder literally exactly the way users do. Just that going beyond that to try to replicate the feeling of where the money ends up — that’s just more work than it’s worth. It’s not ideal, but not exactly damaging to not do that. Some superfans feel a real compulsion to derail enough to be contributing fairly to Beeminder (♥) and that’s a similar kind of sting dilution and I think the answer is that we just accept that the sting isn’t pure sting and it’s ok, it still works.


Madge Castle also disagrees with us on this and has the following prudent proposal:

What about mitigating the logistical burden by just automatically applying the bounties as premium credit for people?

That almost feels worse, since that’s a classic corporate way to pretend to be paying something when you’re not really. Especially when premium credit is worthless to lifetime folks. We could make a new kind of credit — derail credit — but that’s a can of worms. [UPDATE 20024: We sure did open that can of worms.]

Bottom line, the whole dogfood bounty thing is so fraught and messy in so many ways that I feel pretty fine about killing it, despite the wistful little pang.

PPS: But we’re keeping it for our big meta goals: User-Visible Improvements, the blog, and infrastructure improvements.


Imagine how good a manufacturer of dogfood would have to think their dogfood is to eat it themselves! When a company uses the thing they make, especially as part of running the company, it’s called dogfooding.

We’ve been dogfooding Beeminder like maniacs for years, and paying users when we derail on a set of goals related to beeminding Beeminder. And as of right now, we’re getting markedly more maniacal.

Namely, we — the FTBees [1] and some of the other workerbees — are going to cough up when we derail on any goal, Beeminder-related or personal, plus all the beeminder.com/meta goals [2].

“We’re going to cough up when we derail on any goal, Beeminder-related or personal”

And you don’t have to race to catch it first anymore. Now, when any of us derails legitimately, we’re automatically picking a Beeminder premium subscriber at random and sending them an email letting them know that we owe them money! (You’ll be able to cash in via PayPal, credit for more premium subscription time or, if you’re in the US, via check.) We’ll also be posting the derailment payments to our Twitter account (@bmndr) for all to see.

Our main dogfood goal, User-Visible Improvements, has been working smashingly for us since the beginning, and has kept us on the hook since the graph was created in 2011, pushing us to produce 3,156 UVIs and counting. (And making us cough up a full $1k at once on our UVI goal back in 2013. It’s a pretty good story)!

Stay tuned as we continue to complete our Most Important Tasks, eliminate bad habits, and do more pushups. And maybe you’ll be the recipient of our occasional akratic missteps along the way!



[1] FTBees are full-time workerbees, in case that pun was too subtle. We have 3 of those currently, including the founders, and roughly 4 FTEs total.

[2] Some of the meta goals aren’t really things we’re beeminding exactly, just things we’re tracking. Like number of users, engagement metrics, etc. Other things, like blog posts and UVIs and things we’re doing to make Beeminder grow, are in fact high-stakes Beeminder goals. You can tell which are which by whether there’s a pledge on the goal.