Bruce Daisley is the VP of EMEA at Twitter. We discuss how to innovate in spite of perceived failure and protest.
Twitter is now 11 years old, and has achieved incredible success. But in this position, it has found itself in the centre of the “telecoms or publishers” row regarding how platforms should handle undesirable content. A “tech backlash” in 2017 (albeit one which doesn’t turn up in attitude surveys so much as newspapers) has taken aim at Twitter among other platforms.
In the midst of this, Twitter must continue to innovate. As an internet company with an algorithmic feed, Twitter has a sea of data to help it set its course. But data can function simply as noise.
I wanted to understand how to maintain clarity of vision in this context. How does a company retain confidence in an innovation in the face of criticism? How does it measure success and respond to failure? And can legacy players still learn from Silicon Valley platforms?
Cultures of Productivity
I started by telling Bruce that culture seemed to be one of the main constituent elements of digital failure for our respondents. Did that sound right to him?
‘Yes,’ Bruce answered. ‘I can completely understand it. I spent a lot of time this year looking at productivity stats, and what some people might be superficially aware of is that UK productivity has been growing at its lowest level for 200 years; way slower than France, Germany, the US, and one of the reasons is that the UK is more exposed to service industries.’
‘Effectively, service industries have lower growth in productivity, and I think to a large extent we’re not fully asking ourselves about the tools we’re using in those industries. McKinsey said a third of office time is spent on email – I think that’s an underestimation.’
‘So when we look at that we think we don’t treat the email tool that we use as a productivity tool... I think the way we try to integrate businesses together probably is the most important thing – and is also probably why things go wrong.’
Many tech companies have ostentatious markers of culture nowadays, I said. The sleeping pods, the table tennis, the jacuzzi, the architecture...
‘I think that can be a red herring’, said Bruce. ‘I actually tracked down and spoke to a guy called Bjarke Ingels who is to many people’s estimation the number one architect in the world.’
‘I wanted to ask him about his philosophy about open plan spaces. And he felt that, teams should self organise but should be restricted in size. This idea that you’ve got 1000 people on a floor plate isn’t suitable for productive working…’
Bruce may have been referring in part to the new Apple “Campus”, a circular, white, spaceship-like building, seemingly built with the same artistic style guide used for designing a Macbook. The building is totally open plan - which makes for an impressive look, but might be difficult for individual employees wanting to get on with their jobs. It was finished in 2017.
‘There’s work by a woman called Lesley Purlow,’ Bruce tells me, in which she researched what factors went into people feeling as though they’d had “a good day”, at work. ‘People felt like they’d had a good day at work when they’d accomplished something meaningful,’ said Bruce. ‘We tend to not accomplish meaningful things when we’re in open plan spaces, so I think for me, the slides, the massage rooms, the ball pits, are all a big red herring. It’s about allowing people to get a degree of concentration to get work done.’
Cultures of Innovation
Culture is determined by more than just physical architecture, of course. I wondered what Bruce thought about business philosophies concerning failure. To encourage innovation, a great deal of literature seeks to combat natural corporate aversion to failure. Employees should be allowed to try things, and if they fail, it should not harm their careers, CEOs are told. In start-ups, one well-known mantra is that a company should “fail a lot - and quickly.”
‘Eric Ries is a big start-up icon; one of the things that he talks about is trying to get a minimum viable product set up as quickly as possible.’
Since Ries was innovative and zeitgeisty, companies from older industries would occasionally hire him, even when they were sceptical. According to Bruce, one company said, “a minimal viable product for a motor takes several years to produce.” And his response was, “I wonder if we can make a proof of concept which is far quicker than that.”
Bruce broadly agreed with the wisdom. He pointed to the new “Amazon Echo Show”. It’s the first time that an item from the Echo product line has a screen on it, which is designed to mitigate some of the problems with the screenless speakers. But, as Bruce said, it feels ‘immediately dissatisfying’ to use. And you know that by the end of next year, that will have evolved substantially… I think spending too long getting a product to market, rather than adapting it once it’s in market, is probably a mistake.’
I point out that Jeff Bezos draws a distinction between “reversible” and “irreversible” decisions, and insists on making the first type as quickly as possible for rapid failure. The second type goes through a much longer system of checks and balances. Bruce nodded his head.
‘For me, that Bezos model is the one that a lot of people have been talking about this year and the one that’s generated a lot of interest. Of all the Silicon Valley companies, even though Amazon is fourth or fifth in the world, most people would probably bet that Amazon would be the number one company at some point, because their execution is so accomplished and the way they integrate things is so seamless.’
‘They appear to have very few big missteps, but they have a lot of small things that don’t quite work.’
Is it best to invest in a new tech (such as Deep Learning) straight away? Or to watch what your competitors do and then try to implement a version which learns from their mistakes?
‘I think there’s a lot of evidence that “fast second” is the best place to be. But if you’re working on stuff, getting it out there and in front of people is far better than waiting for some big reveal.’
Cultures of responding to feedback
The 280-character limit was doubled this year from 140. This may not sound important, but the short, birdsong nature of the blog fragments on Twitter is arguably their USP. Some would contend that 140 is the difference between Twitter and say, Facebook.
As a consequence, there was a huge backlash to the change among prominent Twitterati. The founder of Angellist even announced that they would only tweet <140 (and certainly only retweet <140) in protest, because “280 is prose; 140 is poetry”. But the change rolled on. And the results have been largely positive.
I wondered how to know when to hold your nerve, and when to respond to feedback. That could be human criticism, or it could be data-based criticism, where a change causes an initial drop in revenue for example, for an anticipated long-term gain.
After all, Patreon recently tweaked its business model. Then, after receiving criticism, they decided to un-tweak it, even issuing an apology on their website. How do you know when to U-turn?
Bruce thought about this. ‘Almost all of the innovation that Twitter’s ever done has been user-led. So we’ve just launched Tweetstorms and that’s largely the ability to post multiple tweets strung together. And that’s because people were doing it. And hashtags, that’s because people were doing them, so they became a clickable thing. With usernames, somebody just said, why don’t I put someone’s @ name before their username, and it’ll show their name. Retweets were created by users.’
‘With the character limit, what we were finding was that a whole load of tweets were being abandoned at 139, 140 characters ... and people were leaving. But when we looked at non-English scripts – like Arabic, Japanese, we were finding that they weren’t having the same abandonment. In Japanese you can say about a paragraph in 140 characters. And we were finding that people weren’t abandoning. Twitter is the number one social media site in Japan.’
‘The number one bit of user research we do is we ask what do you like about Twitter what do you dislike about Twitter... the thing that comes back all the time is when people tell you, “I don’t know what to tweet, I find it hard to tweet” so it was to try to answer that. It was a data-led response.’
‘I used to work at Google and the one thing we knew there was that every change that we made, the stuff people were the most unhappy about when it first happened, was the stuff where when we offered to take it away from them, they were most mortified by.’
‘The Google search bar used to be... just a line of text. It was tiny. You were effectively writing in 12 point font. Then they made it huge. And people immediately said, this looks like a children’s toy. This is like the Fischer Price version of Google. It looks ridiculous.’
‘We gave people that for three months. Then when we said, “we’re going to go back to the old one”, they were really upset. Because with any social product, because people spend so much time on it, it’s like changing your mum’s face. It’s something you’re so familiar with you feel really unhappy about it. So what we find is that any change provokes this annoyed dissonance.’
‘But broadly in the spirit of all innovation a lot of our innovation here comes from, a couple of times a year we do something called “hack week”. And that’s where someone from anyone in the company, in sales, finance, design, can come up with an idea, and then effectively, builds a tribe to try and build it with them. So, they come up with a minimal viable product in a week. Then we have a big craft fair where people take tables and try to win votes from their project. Most of our big innovations in the last two or three years have come from that.
Metric and identity-based decision-making
A company like Twitter poses the question, how do you measure your ambition in the first place? How do you even know what success or failure looks like with nobody analogous to benchmark against?
‘We’ve got 330 million who use Twitter. But maybe there’s not one billion people who want to build a timeline. Maybe there’s a product that should magically build a timeline for you or should put things in front of you that are interesting. So we’re always looking at the limitations of our approach.’
As for Twitter versus Facebook, Bruce says, ‘we see the use case differently. The thing we’ve tried to do in the last two years is define in a reductive sense what Twitter is. There’s a great bit in Nicholas Carlson’s “Marissa Mayer and the fight to save Yahoo.” It’s a book about the story of Yahoo. He talks about how when Carol Bartz was the previous Chief Exec of Yahoo, she took 20 people out on an off-site and she asked them all to write down a single word for Facebook, for ebay, for Google, and for Yahoo. And of course, they all wrote down “friends” for Facebook, “auctions” for ebay, “search” for Google, and there was 20 different words for Yahoo.’
‘I think what we’ve been able to do at Twitter since Jack Dorsey’s been back is to ask well what is the single word for Twitter. And Twitter is basically – the single word’s probably news, but Twitter’s about what’s happening right now.’
‘We’re very metrics-focused, and I think the critical thing is to follow the right metrics. So for example we said to Wall Street that the metric we were going to follow was Monthly Active Users (MAU). And we pretty quickly realised after doing that, that more important than a MAU is a Daily Active User.
‘What you tend to find with MAU is that [they use Twitter] once every two months. ...They’re immensely low value. They’re turning up, checking what happened to that train, and leaving. When you engineer for that, you’re very quickly just telling people to come back out of the blue. You send them an email.’
‘When you’re trying to get a DAU, you’re trying to build a product that’s useful, habitual, that’s delivering on a very demanding set of criteria. We normally open three or four apps a day. So if you’re going to get into that list of three or four apps a day, you need to be really good for them to think, “I’m choosing you”.
‘Being metric-driven is one thing, but getting to the right metric is more important. For us, optimising for MAU, our published number, was actually limiting us. Because it wasn’t forcing us to ask the hard questions – like “what are we actually building that’s useful for people?”’
‘It becomes exhausting if you’ve got a dashboard of 40-50 numbers that you’re trying to measure. And what we’ve tried to say is that we’re looking at two numbers. And it's these two.’
‘Ok we’ve actually said it’s these two numbers, and you get rid of the noise, it can be more useful.’
1. Innovating through noise, with the VP of EMEA for Twitter, was first published in Nimbus Ninety's Digital Trends Report 2018.