Experimentation has been a critical topic in product management over the last several years. However, an interesting trend I’ve observed in 2023 is that of taking experiments much further than in years past – and shipping some seriously half-baked products that were positioned to the market as being generally available. This is not coming from startup organizations, but big technology itself, which makes the trend even more curious.
A first case study is that of Microsoft and the venerable tool known as Outlook. Several years ago, Microsoft began what is known as Project Monarch to modernize and unify the desktop Outlook client across platforms. This became known as “Try the new Outlook”, first to macOS users and then later Windows users. Under the guise of “try” I have no criticism of Microsoft’s actions. However, in recent Windows 11 builds – Microsoft has begun replacing the Mail and Calendar clients debuted with Windows 10 with the new Outlook. And a number of key features are missing without clear equivalence:
- Support for non-Outlook/Hotmail or Microsoft 365 accounts (later remedied in recent builds to include GMail, Yahoo, and IMAP finally)
- Offline support (coming in a future release) or PST files
- Built-in tasks support (opens Microsoft ToDo on the Web, not even the desktop application)
- Ability to switch sending accounts while preserving replies (this has worked forever in historical Windows releases as well as the macOS version of theoretically the same codebase)
- Any sort of touch-friendly UI, which was predominant in the Windows 10/11 Mail & Calendar applications (Microsoft has seemingly abandoned touch with Windows 11 as a scenario)
And…there is no clear roadmap nor level set to user expectations as to how these capabilities will be addressed. Fortunately, the distribution surface area of this appears to be quite limited and IT pros have the ability to block the new Outlook from replacing the old Outlook. Hence, minimal market impact (for now).
A second and more interesting case study is that of Threads from Meta. Clearly this was launched very early to help capture the user exodus from X/Twitter given the multitude of changes in that social network. This was positioned effectively as a generally available social network. A number of key features were missing and added later; some are still missing and in the roadmap. Examples:
- Android client (added quickly)
- Web Client (added later)
- Hashtags (semi-supported)
- Lists
- GIFs (added later)
- Polls (added later)
- Direct Messages (still missing)
Assuming the intent of this release was to capture consumers, it can be easily excused that business features are missing such as:
- APIs for Business (still missing)
- Analytics
- Etc, etc.
Given the breadth of this launch, it definitely has had customer and market impact. Usage quickly grew to over 100M active users, and then plummeted precipitously – making the statistic more like active downloads. Active users is now climbing again – and in Meta’s most recent earnings call, CEO Mark Zuckerberg made it known that the network is on track to have over 100M active users once again. Part of this for sure is aligning to Meta’s narrative that they are adding features. But, my anecdotal observation is that usage is equally motivated by users simply upset with X/Twitter and migrating given lack of viable alternatives. Threads will reach its goals, albeit at X/Twitter’s expense. Which is likely Meta’s business strategy, just not publicly articulated.
I do want to give Instagram chief Adam Mosseri HUGE kudos for being incredibly transparent to the entire user community and openly solicitous of user feedback. He is a case study in how to lead in a situation of this nature.
I can’t see such launches happening from big technology in years prior. Normally, one cannot do wrong by following in their footsteps. I would advocate shipping below minimum viable products or features such as this, however, is not a practice one should contemplate following – in this form… Big technology has the resources and wherewithal to withstand any blowback (especially if it is calculated, such as capitalizing upon frustrated Twitter/X users); a privately held company may and probably does not.
I think a reasonable middle-ground that has played well in years past can continue to serve well – make experiments or less than fully baked features or products opt-in, with a clear Beta or Preview label. Nobody is advocating a 5-year beta like GMail (2004->2009), but I think clear level sets with users are what will preserve and further brand trust while still offering considerable product flexibility to try out new and/or incomplete ideas.