Three Hypotheses On Collaboration, Hypothesis 3: Asynchronous Communication is Critically Important and Broken

Drew Dillon
ProductMan
Published in
5 min readNov 7, 2017

Over the course of my career, I’ve developed three foundational hypotheses about unaddressed needs in collaboration:

  1. Knowledge Evolves
  2. Email Inboxes are the Intranet
  3. Asynchronous Communication is Critically Important and Broken (this post)

In this series, I will explore each of these hypotheses and some of their implications.

Asynchronous Communication is Critically Important and Broken

In the age of Slack, fellow tech bubble dwellers might be forgiven for thinking it’s the alpha and omega of all communication needs. I’ve discussed the reasons for asynchronous communication elsewhere, but, in summary, asynchronous communication is best when:

  • Realtime isn’t needed or expected — a lot of reasons for this, remote offices, external clients where social pressure to respond quickly might be rude, that one idea hits you off hours, etc.
  • A concept needs to be deeply considered — chat threading, and especially the lack thereof, make it exceptionally hard to deeply discuss an issue. Further, chat structure prevents the creation of a legible searchable archive of decisions.
  • You don’t know the other party well — exchanging an email address is still the lowest friction way to get in touch with someone.
  • Broadcast — addressed messages are tasks and the inbox is the message task manager.
  • Addressing multiple constituencies — if I have a CS channel and an Engineering channel and I need to send the message to both groups. Do I send it twice? Create a new temporary group? Post it for the whole company and annoy people who aren’t in those groups?

I could go on, but, suffice to say, I believe asynchronous communication to be a vital medium of communication. But it’s broken, that’s the name of the hypothesis, right?

Story Time

Let’s rewind the clock a bit. 2007, David Sacks was CEO of genealogy site Geni. Paypal’s culture had been famously anti-meeting and pro-email. Sacks, having contributed much to that culture, believed in email as a way of reducing time to make decisions and distribute work.

What Sacks didn’t love, though, were all the byzantine rules surrounding email distribution. As CEO, he couldn’t always know who was the best person to answer a question or perform a task.

email would be so fun if everyone were fused at the torso

He would send messages to his best guesses. Email would travel these strange paths of distribution, replying and forwarding through an ever-more complicated tree. Sometimes he’d get an answer back and sometimes he’d have to poke again, etc.

Twitter had just come out and captured the tech community’s imagination. Sacks saw the Twitter feed as a solution to his problem and segmented part of Geni’s workforce to build the first version of Yammer.

Twitter is synchronous, however, like a giant chat room. So, as Yammer grew, it became incredibly noisy. Conversations would flow and those answers/assignments would get lost in a reverse chronological feed.

This lead to Yammer adding threading, looking more Facebook-like, moving ever further from a following model to groups. But this created it’s own problem, the single central feed lost the benefit of serendipity. Engineering saw what Customer Success cared about, Sales saw discussion around the features they’d asked for, etc. And those times when Sacks’s questions might not find the right people, there would be a critical mass of everyone to route the message.

The central feed needed to be algorithmic, lest users be shoved off back into their organizational silos. So that’s what we did.

Still Broken

I wish I could say that the $1.2 billion acquisition of Yammer meant we fixed the problem. We missed something, which I’ll discuss at the end. Email still rules asynchronous communication and is still broken for those reasons:

  • Addressing is complicated — who you send to is just as important as what you send. Things like FWDing and BCCing split the tree, losing context and not benefiting the initial sender.
  • Serendipity is zero — only the people who get the message can act on it or learn from it.

Ultimately, though used daily by millions of people, Yammer wasn’t able to resolve the issues and fix asymmetric communication. Within the reason for this lies the solution.

Email-centric Collaboration

Yammer would, on occasion, be dinged for low engagement. Of course, commentators don’t know how hard it is to build a social network. My guess is that if you launched 50,000 Facebooks in parallel dimensions, a majority wouldn’t work out.

The real answer was more difficult: email.

All it took for a dev team to abandon Yammer was a single engineer refusing to leave email. You might’ve read my series on Yammer’s initiative model, the thrust of three months of 30+ people’s time was to work on this exact problem.

My final takeaway was that any improvement in asynchronous communication needed to start with email. When I say “email”, I’m generalizing a few things:

  • Servers — big machines that speak to each other to route messages around the world.
  • Clients — apps, sites, etc. that make email digestible for users
  • Communication Protocols — SMTP, how email is sent around the world, and IMAP, how disparate clients interact with email servers.
  • Messages — effectively web pages as architected in 1999, but only visible to addressees

If you want to fix asynchronous communication, you have to start at the protocol. SMTP is an old standard, but it’s ubiquitous and extensible. SMTP could be extended in such a way that a new server could understand different delivery rules. Those enhanced delivery rules could provide a virtual central feed, enabling serendipity.

Replacing servers is hard, though. The goal here isn’t to use a decades-old protocol, but to avoid the in-group out-group problem that happens with all collaboration tools.

If I were to build Yammer today, I would start with SMTP.

Benefits Beyond Routing

You might recall that email inboxes are the intranet and that all artifacts of workflow are replicated in email messages. The holy grail is an intelligent server/client that reads those update email messages from Salesforce, Jira, Expensify, and turns them into actionable intelligence.

All of your business systems integrated, no APIs required.

So What?

We’ve all worked with that one person who seems to just route email around the company. I’m here to tell you that’s true and that that routing is a vital part of your organization. They’re mechanical turking around addressing rules and lack of serendipity.

Stagnant communication is tantamount to thermodynamic equilibrium. And a company where information doesn’t get routed around by some intelligent actor faces the equivalent of heat death.

Asynchronous communication has to evolve for this century, but, rooted in the past, guide everyone to that better place.

Read the rest of the series:

  1. Knowledge Evolves
  2. Email Inboxes are the Intranet
  3. Asynchronous Communication is Vital and Broken

⬇ Follow me for more on Productivity and Product Management.

--

--