Testing in the Dark: Lessons in Cross-Site Communication
The third Midlands Exploratory Workshop on Testing (MEWT) took place in Nottingham this weekend, themed around the subject of communication. Each invited attendee is given a 15-minute speaking slot to give their own spin on the topic.
Sadly I wasn't able to attend on the day, but I'd already produced my slide deck so it seems a shame to waste it. Here, instead, is the same presentation in essay format! (The original slides are available on Slideshare.)
When you can't see your colleagues
I've worked for a number of large multinational organisations with development spread across multiple sites worldwide. Even when working within smaller local teams, I've had colleagues who are remote workers, or have occasionally telecommuted myself. This introduces a number of communication challenges, most of which - at first glance - appear to be solely the result of the lack of visible nuance.
Some of the problems that I've experienced include:
- Misunderstood messages: "Crossed wires" are more likely to occur if you can't get immediate feedback/clarification at the point of receipt. Similarly, if you're reading my email, I won't see the confused look on your face at a key moment of misunderstanding.
- Is anybody there?: When you're sending an email or instant message, then (regardless of things like status indicators) you can never be sure whether somebody has actually received your message and is silently stewing on it, or whether they're in a meeting.
- Unpredictable response time: Similarly, if you're having a face-to-face communication, you can generally expect an immediate response to a question (even if that response is "I don't know, I'll have to find out"). On the other hand, with remote communication, even the simplest questions have a lag time which turn them into long drawn-out exchanges.
These result in, for me, a much less satisfactory communication experience. It's rarely the fault of either participant; it's a nature of the medium. You can mitigate with telephone conversations and webcams, but these have a higher interaction cost which some find off-putting: the same people who will happily turn and chat with a colleague on the next desk will balk at the suggestion of picking up the telephone and speaking with someone in another office. Many people find comfort in their geographic position, using it as a wall to avoid interruption, when it can actually result in bigger inefficiencies further down the line.
But words are still important!
In researching for this presentation, I planned to cite a popular statistic about how the words within communication are much less important than how you look and sound. It's an assertion that I first became aware of during Eddie Izzard's 1998 stand-up show Dress to Kill:
However, digging deeper, I discovered this was a "Leprechaun of Communication" (hat-tip to Lauren Bossavit and his excellent Leprechauns of Software Engineering). The original research on this subject (Albert Mehrabian, 1971) was much more specific: it dealt with communicating single-word expressions of emotion, and how incongruences can create "mixed messages" (for instance, somebody saying that they're "happy" with a sullen face). As the years have passed, his findings (which had their own statistical flaws in the first place) have been repeatedly over-generalised, such that the mythical "7%-38%-55% rule" is now often applied to all communication.
This over-generalisation can be very easily debunked:
- If we only get 7% of our understanding from a person's words, shouldn't we be able to understand 93% of what a foreign speaker is trying to communicate?
- If body language alone can account for 55% of all communication, shouldn't we be able to understand more than half of a televised speech if it's on mute?
- Critically, as inefficient as instant messaging can be, is it really only 7% as efficient as a face-to-face conversation? Anecdotally, this seems far too low.
So if words matter, maybe there's more to cross-site (mis)communication than simply not being able to see the other party.
My gripes with (and solutions for) email and instant messaging
When we resort to text-only communication, we often try to counteract the lack of visual in a variety of ways. However, often these compensation techniques have their own side-effects. For instance, what springs to mind when you think of that one person in your office who still uses the "read receipts" feature in Outlook? Or takes a one-on-one email communication and forwards it to a 50-person distribution list that rumbles on forever?
Then there's emoticons. Their informality seems suited to quick internal back-and-forths, and (in lieu of your actual face!) can communicate some kind of visual expression, but I've seen them cause weird confusion. For instance, I once saw a scenario where a person used the :/ "annoyed" emoticon repeatedly, when they were actually trying to communicate a wry smile. You've probably also experienced the issue of receiving an email with a random "J" in it, where the J was actually an Office smiley face symbol (in Wingdings) which gets corrupted in HTML emails. These are situations where we're trying to add meaning, but often only add confusion.
In order to combat some of these problems, I've adapted a few strategies:
- Avoid email and IM wherever possible: Go real-time wherever possible. This should theoretically be easy if you have a telephone and/or a webcam, but getting teams comfortable with this can take time.
- Adopt simple usage rules: When I can't have a live vocal conversation, my general rule is to use IM for "one-liners" (simple questions with straightforward answers), and to use email for anything that can't be represented in an IM; however I'll still try to stick to "one-point-per-thread".
- Adopt "Inbox Zero": I treat my inbox root as a To Do list. If your mail is still there, it's because it's immediately important to me. Anything else gets filed into folders (often using automated rules) or ends up in a catch-all "Archive" folder when I'm done with it. I do this with my personal and work emails, and (much like a Kanban WIP limit) if there's more than about five emails, I'll start to push back on things.
- Give context regarding urgency (or lack of it): If you need something soon, say so. If you don't, say so. This includes proper and responsible use of the "Urgent" flag in Outlook.
- Create information radiators: If you're often exchanging simple questions-and-answers on the same subject (e.g. "is build X ready for test?" / "is feature X due to be completed soon?") then consider surfacing this information via dashboards or wiki pages. Do it in a format where it's always up-to-date whilst remaining low-maintenance, so that it can be relied upon but doesn't need much of your time.
Effective and efficient meetings
So far, most of the interactions that I've talked about have been specifically regarding one-to-one communication. However, as soon as you need multiple heads to make a decision, thoughts rapidly turn towards the conference call. These are generally flawed beasts, with the format being significantly less satisfactory than a face-to-face meeting would be, as brilliantly summed-up in this video:
And not only do meetings become less efficient as more people (or more locations) are added, but the cost of the meeting rises too. I've found that a quick way to put a kibosh on out-of-control meetings is liberal use of this Meeting Ticker, which (given a number of attendees and an approximation of their hourly rate) shows a real-time ticker of what the meeting is costing your company. Before long, people look for smarter ways to find the same information in a less costly arena.
There's a rising number of great one-stop productivity tools which look to lower the boundaries to group communication. HipChat, Yammer and Slack are the best of the market leaders that I've used, although there's an intriguing wildcard: Sqwiggle.
Sqwiggle's USP is its grid of always-on, frequently-updating webcam snapshots which allow you to see whether someone's available without relying on traffic lights, and see whether they look particularly busy or stressed. You can also tap thumbnail(s) to immediately start a video chat with one or more people. It's the best substitute of face-to-face communication that I've seen, though it takes some getting used to ("What, you mean people can see my face all day?") and if a team is fundamentally opposed to the idea then it won't ever gain any traction within that team. I've seen successes with it though, and it's surely worth a trial; if the worst that happens is that the team finds new ways to express what works (and doesn't work) for their remote communication, then that's still progress for you.
There's an excellent essay by Janice Nadler and Donna Shetowsky: "Negotiation, Information Technology, and the Problem of the Faceless Other" (PDF link) . I'd definitely recommend this as your next step if you're experiencing the same challenges as me. Alternatively, leave a comment below and I'd be very happy (smiling, upward intonation and positive language!) to debate further on the subject.