Advance warning: this is a bit of a rant, but please take it as “tough love” and constructive criticism rather than pure snark.
What’s going on with interactive graphs? I first heard about them in the weekly beemail almost exactly four years ago (10 Jan 2019):
12 days later, another beemail said this:
That was supposed to be in an Oprah voice. Anyway, we’re gradually getting these new graphs turned on for everyone and if you’re reading this (daily and weekly beemail subscribers) then all your graphs are currently being generated by the new (and open source) javascript graphing engine.
Please let us know of any differences you notice, good or bad!
Next steps (all theoretically quite easy now):
- Turn this on for everyone and officially retire the old Python code.
- Switch to SVG images which should look even better.
- Switch to interactive graphs!
Again, we’re eager for more feedback on the new graphs before we complete the switch. Just hit reply on this email! Thanks everyone!
I remember getting quite hopeful at the time when reading the “all theoretically quite easy now” bit. “Ah cool,” I thought, “it must be coming soon!”
Then in a private email discussion with @dreev in Sept 2020 (i.e. 20 months later):
and the prototype version of interactive graphs – road.beeminder.com – are getting closer to ready to being part of beeminder proper!
In Oct 2021 he posted in this forum that:
at the moment we’re more focused on things like merging graph.beeminder.com into beeminder-proper
and then another post in Jan 2022 saying:
Even dumber is that we’re still paywalling the static graph editor even though graph.beeminder.com is free to all.
This gave me the clue I was looking for, and finally I found premium – beeminder where it shows that the “graph editor” feature is available for Bee Plus and Beemium users. But https://graph.beeminder.com/ still works fine, so I guess this inconsistency hasn’t been addressed since it was discovered 12 months ago?
If you want to incentivize people to pay for premium features, I’d suggest making it more obvious what those features are. For example, you could have some text next to the static graphs telling users that they can get better graphs by upgrading. You could even disable the free graphs 
Although IMHO that would be a mistake. The static graphs are pretty horrible without the ability to zoom, so I think it would make more sense to ditch them altogether (or at least offer them as a fallback for NoJS fans). Instead you could provide a zoomable interactive graph by default, and then maybe limit some of the extra features (such as editing) to premium users if necessary. It feels wrong to me to have a business model where users have to pay to upgrade from a poor UX to a good one.
But actually, it seems to get quite a bit crazier than that - I also found a link from the Premium page FAQ to Graph Editor - Beeminder Help which says:
There’s no “preview changes” functionality, so you can instantly derail your goal if you make a bad change. The Undo Graph Edit button will restore your graph to its previous state, and you’ll need to reply to the legitimacy check for support to cancel the charge. You’ll probably prefer to use the visual graph editor instead, which will not allow you to submit a bad change, allows previews, and allows you to drag the lines around rather than having to figure out the row system!
Let me check if I understand this right. You have a graph editor as a premium (paid) feature, which in at least one respect works worse than a freely available alternative which the official FAQ recommends using over the premium feature? 
I don’t want to say I have a love/hate relationship with Beeminder, since “hate” would be far too strong a word. It’s more like love tinged with sadness and disappointment. I so want to love it unconditionally, since the concept is great, and it has so many great aspects… but my experience goes in cycles: I use it for a while, then one or two things get so irritating that I ditch it, only to come back a year later when I’m fed up of not maintaining some key goals (mainly inbox zero).
Last time it was the unreliability of Android notifications which pushed me away. Yet here I am back again 18 months later, and to be honest the site doesn’t look significantly different from what I remember in 2019, let alone from 18 months ago. The clunky static graphs are a big part of that impression, but in general it feels to me like the whole UI/UX isn’t doing justice to the quality of the service, and deserves a redesign.
I noticed several other issues whilst restarting my gmailzero goal. For example, the default slope was something crazy like 0.0000014 mails per day. Why would I ever want a goal to increase the number of emails in my inbox?! And why present the number in such a user-unfriendly way? If it’s such a slight slope, why not default to measuring it in mails per month or year, to reduce the number of zeroes? Or even better, just start with the actual goal number rather than the gradient.
I also notice it falls foul of the dreaded Javascript floating point precision issue, e.g. it was on -4.6 and clicking the minus button yields:
These are all minor things, but it’s a case of death by a thousand cuts - they all add up and really impact the overall feeling of quality. Given your admirable public commitment goal to continual user-visible improvements, I can’t really understand why issues like this weren’t tackled a long time ago.
Since this has been a largely negative post, I hope I can lighten the mood by sharing the greatest tech lightning talk of all time, in case you didn’t see it before. Or if you did, here’s a couple of otters holding hands:
Either way I hope this feedback is useful!