[notes] Feedback on Walletbeat from Vitalik

[polymutex]

Today I got to sit down with Vitalik and ask for feedback on the site. He opened it up on his phone and gave some thoughts as he browsed the site. I am using this issue to recall the interaction and track the feedback I can remember somewhere before it escapes my memory; we should consider filing individual bugs for these.

Getting the site to open:

• First website that popped up was walletbeat.xyz which does not exist. It looks like it's on sale for 399 USD, which is not great but might be worth stomaching and having as a redirect.
• Then he browsed to walletbeat.eth.limo, which didn't work because only the beta. subdomain is set up. We should set up a redirect from walletbeat.eth to beta.walletbeat.eth.
• On beta.walletbeat.eth.limo, as he was on his phone the layout was not very usable. Need better mobile UX in general. The left part of the table (wallet title etc) covers up most of the usable browsable surface.
• [my feedback, not his]: On mobile, more than half of the above-the-fold space should be wallet data (past table headers). Current logo and table headers etc are too tall for phone layout.
• Vitalik has a phone that can unfold into a dual-screen square-ish format, so we used that and the "Request desktop site" feature of the browser to get a better layout.

UX feedback on comparison table page

• First piece of feedback was that this is too many pie charts. I pointed to the stage system but I think the thing to address here is that the pie charts are more visually prominent than the stage label. If we want the stages to be what users focus on by default, then the stage label needs to be what eyes would naturally focus on. Perhaps that means making the pie charts smaller, perhaps it means replacing them with a synthesized version (on the comparison table only, not on per-wallet pages), perhaps it means replacing the 5 pie charts with a single one that has 5 slices on it representing each attribute group. And maybe make the stage system label more prominent somehow as well. TBD.
• Vitalik tapped through each of the pie chart slices one at a time. On mobile this causes very confusing behavior, because one tap causes the "hover" effect (showing the tooltip), but the next tap causes the click event (which unfolds the wallet row to expanded view). This is compounded that mis-taps are likely; sometimes he accidentally re-tapped on an existing slice which was confusing because it made the wallet row jump and cause his intended target slice to jump with that; other times he tapped on the outside-of-pie-chart region of the pie chart table cell which causes the wallet row to un-expand and it's not clear why because it feels like tapping a slice. Overall I think the feedback here is: forget about row expandability and hover effects on mobile. It's just not gonna work due to the limitations of touchscreens and the difficulty of touch targets.
• The stage system label toolttip wasn't really readable (can't remember the details). I'm not sure what to do about this, maybe on mobile it should just go to the per-wallet page.
• Clicking on the non-title-text but title-cell of the wallet causes row expansion, but it is also very easy to mis-click. On mobile (and perhaps on desktop too?) the whole non-otherwise-interactable-area should probably go to the wallet page.
• V accidentally clicked on the sort button and asked what score is being used; I described the internal weighted score we have but that this would be changed to the stage system once not every wallet is stage 0.

Once on the wallet page, he gave feedback on stage methodology.

Per-wallet page stage methodology

• V initially suggested that the stage system should be structured such that "at least 10% of wallets would not be stage 0", otherwise it is too high of a bar and discouraging, and having wallets as stage 1 would act as feasibility proof.
• V then looked through the criteria for stage 0/1 (not 2); feedback there was that all makes sense, except the ENS attribute which should not require chain-specific addresses because no standard for that has yet been settled on (and he's right). The other two preventing Ambire from reaching stage 1 are light client verification and private transfers which he seemed to agree with, but perhaps this means we cannot have any stage 1 wallet at launch unless they ship these two features before launch time.

[maykelxyz]

perhaps it means replacing the 5 pie charts with a single one that has 5 slices on it representing each attribute group.

Agree on this, it might be too much for a user to be looking at 5 diff pie charts on the get-go. Maybe we can show the 5 different pie charts on the wallet page and refactor the dashboard to contain a single one with 5 slices on it representing each attribute group as mentioned in the feedback.