Classic mistake resulting in low growth!

Oh boy, I have a story to tell. So, I refactored the signup / subscribe flow for Minecraft Playdates right after I launched because I realized that if you selected a plan before you had created an account, that by the time you finished with account creation, I had lost the state of which plan you wanted to subscribe to. (The analytics showed huge dropoff on the signup -> subscribe flow).

So... Lesson <foreshadowing>: Test everything. I smoke tested, and everything seemed to be working, so I pushed the new site live... A few days later after letting it soak, turns out that this fix had the opposite impact, it increased the dropoff rate rather than decreasing it. Hmmm...

Well, purely by chance, through non-scalable growth hacking techniques (reading Minecraft for Parents groups on Facebook, posting helpful comments for parents having trouble getting playdates setup for their kids, etc), I figured, why not try attracting early adopters via a compensated user study? (details here if you're interested), and quickly got a bite from a parent in one of those groups.

So, getting her setup, she writes to me "Hey, I hit the subscribe button but nothing happens". <queue the dread...>

So we find a good time for a video chat / screenshare (remember our value prop: parents that don't need to know anything about tech to get their kids setup with playdates!), she screenshares with me and shows me the problem, she pops up the javascript console, and there's the dreaded error: "billing.js:437 Uncaught TypeError: Cannot read property 'id' of null"

Gahhhhhhhhhhh

So after our chat, I'm thinking WTF .. as I start debugging. Yeah, you know software dev 101... test your changes... well... yeah. So now I try to go subscribe again for myself, and sure enough, I get the same error!

OMG. I broke the site with my last round of changes.
OMG. I didn't actually retest subscribing after that set of changes. <facepalm>

OK, well, the fix was actually really simple. While refactoring the flow, I had renamed the entity id of the subscribing modal that pops up while the subscription is processing, but didn't update the corresponding Javascript that pops it open.

So I tested in the sandbox, it seems to work. I cautiously push to production, and... Yay, it's working now <wiping sweat off brow>!

You'd think that after writing software for 25+ years, that I would have learned this lesson... Well, turns out you'd be wrong. But truly, I get it now, it'll never happen again <knocks on wood>...

Anyway, this story ends on a positive note. That user, this didn't turn them off! Their kid is having fun playing Minecraft now, and I have my first customer, and in a week or so we'll do a UX feedback session. Yeah!