Release 1.18.0 - 27 November 2020

[paolodamico]

Overall priorities

• Session recording
• Plugins

Detailed contributions

Karl & Paolo

• Session improvements Sessions page UI/UX improvements #2352 Session recording where css had not rendered #2365 Create a custom session recording player #2332
• Person improvements Persons pages UI/UX improvements #2353
• Deep dive retention analysis

Eric / James G

• Retention EPIC: Retention #2228, Retention cohorts graphs #1936
• Lifecycle Lifecycle statistics #1209
• Separate event ingestion from rest of the app
• Move clickhouse tables around for storage (make sure we don't run out of disk)
• Migrate everything to AWS
• Log availability
• Stretch goal: event consumption in Go

Michael & Marius

• Enable plugins on app
• Add unit tests to plugin server
• World-class documentation and dev tools
• Scheduled plugins (fun distributed locking problem?)
• API plugins (webhook -> plugin)

Tim

• Dedicated support/Sentry/random bugfixes/random UX improvements
• Keep a log of everything you've fixed/touched

[paolodamico]

General stuff

• Session improvements (Sessions page UI/UX improvements #2352). Rationale (disclaimer: absolute numbers are still pretty low to be too significant, but should give a general idea):
  • From customer calls, session recording seems to be a game changer, but the experience is still a bit rough.
  • As a general preamble, despite this being a beta feature that requires explicit opt-in, we're seeing steady grow in active users WoW. As additional context on WAUs, this last week we had the same number of WAUs that watched a session recording and created an action.
  • From users that used it on week 1, 60% were retained to week 2. From the users that were not retained in the session recording feature, 50% didn't actually come back to PostHog at all this week.
• Persons improvements (Persons pages UI/UX improvements #2353). Rationale: Related to sessions above. The new UI/UX covers more use cases related to session recording (e.g. understanding what happened when a user experienced an error, or when they completed a purchase, ....)
• Implement toolbar improvements (Toolbar UI/UX improvements #2370). Rationale: We have several reports of bugs (most of them coming from customer calls), and from sales calls we know that users do not understand this feature unless we walk them through it. Perhaps the most important here, we don't have this feature well instrumented to capture enough metrics to be able to make a stronger decision of pushing this forward or not.
• Improve our own tracking (Improve core action tracking #1575). Rationale: there's still a lot of stuff we don't have good metrics on, and it's hard to make strong product decisions without reliable data on everything.
• Retention epic (EPIC: Retention #2228), most important Retention cohorts graphs #1936. Rationale: we really need this to analyze our own data and improve our own product.
• Bug Bug with contains filter in array-based attributes #2301. Rationale: part of core functionality, without it the experimentation power of feature flags essentially disappears (without being able to measure, there's no point in A/B testing).
• Cohorts improvements (Cohorts UI/UX improvements #2350). Don't have a very strong argument for this right now other than the reported pain points of a couple of users and that if we work on Sessions page UI/UX improvements #2352 & Persons pages UI/UX improvements #2353 this will ensure consistency.

Myself

Will only have 4 working days in the next release cycle.

• [UX] Pair up with @macobo to work on improving session recording playback experience (Create a custom session recording player #2332 / Show current URL in session recording, allow jumping to when page was visited. #2129). Rationale: see rationale for improving session recordings above.
• Deep dive retention analysis. Rationale: we need to understand retention a lot better to improve it.
• Update retention docs (Update retention page posthog.com#654). Rationale: in the last week 21% of active users made use of this feature (core feature), we're introducing a ton more functionality, and it'll likely be a source of questions / confusion.
• Activate billing for everyone & make sure people near the event limit are prompted to update.
• Stretch. Improve internal referral experience #2367 / Migrate team signup to React #2336 to increase internal sign ups & indirectly retention. Rationale: hypothesis is that users whose team members are on PostHog retain better.
• Stretch. Fix weekly email reports. Rationale: hypothesis is that this will help retention by driving users back to the product with useful & actionable stuff.
• Stretch; Lower priority. Docs dark version design proposal.

[macobo]

Topics

Session recording

Session recording has gotten a lot of buzz, I think it's time to give it proper attention and make it great.

• Make sure all recording data is reaching us, is valid: Session recording where css had not rendered #2365. Note that this is likely an ongoing thing due to amount of moving pieces.
• Update sessions views to more emphesize session recordings Sessions page UI/UX improvements #2352
• Create a custom session recording playback experience Create a custom session recording player #2332
  • Related tickets: Back N seconds Add "back N seconds" button to session recording player #2131, Skip inactive UX: [session recording] Inactive periods don't get skipped. #2096, Keyboard-awareness: Keyboard-aware session recording player #2363, Better navigation within session recording Show current URL in session recording, allow jumping to when page was visited. #2129
• (Stretch) Make sure we properly delete + allow storing session data on S3: Save old session recordings to S3 #2142

I'd love to work together with @paolodamico on this to have some design brainpower.

Capture endpoint stability

• We're seeing a lot of failures in event capture endpoint. I think I would like to focus on that since that's affecting metrics + could be affecting session recording. (No issue yet)
  • Related: We should set up proper logger + log full payloads and errors if errors occur in backend. Sentry doesn't log full payloads and out payloads are large.

Don't feel too strong of an ownership here, but this seems relatively hard but high-impact work.

Toolbar

• Wrap up suggesting data-* selectors without introducing problems Autosuggest data-* attribute-based selectors over className based selectors #2140
• Showing border around matching elements (carryover from this iteration) When defining actions, show a border around matching elements #2143

[mariusandra]

For me:

• Continue working on the plugins. We finally landed the plugin attachments PR, yet there's still a lot more work to be done
  • Test plugins enough to enable them on EE
  • Scheduled plugins (fun distributed locking problem?)
  • API plugins (webhook -> plugin)
  • Maybe: frontend plugins (new scenes, world map, etc)
  • Plugin dev tools and docs
• And various odds and ends as they come in...

[Twixes]

As for me, one thing I'd like to do is focus on smashing #2128, but happy to take on something else as well – TBD.

[EDsCODE]

I will do as described at the top.

• Retention epic and lifecycle
• General analytics bug swashing and filling out (the sessions analytics could be way more built on because it's very spotty in terms of graphing ability right now)

[mariusandra]

I'd also like to add these issues:

• Expo support for React Native: Feature req: Expo support for React Native library #1174
• Some bug with the RN lib: https://github.com/PostHog/posthog-react-native/issues/1
• Plus the RN timezone issue if it's still an issue: React-native sending incorrect timestamps #2173

[paolodamico]

Closing as this has been released