DX的动态

Ten months on from our first DX snapshot we've concluded our fourth today. At the outset I was nervous about whether the claims around consistent high response rates would be true. There's always doubt about whether it would translate in your own company context. So, I thought it would be worth sharing that our experience upholds the claims. Across a population of around 2,000 developers we've now run four snapshots with each achieving 90%+ participation. But participation isn't the key metric. The outcomes around developer experience and effectiveness that the insight drives are what matters. Our experience is that using this mechanism we've really seen success in teams taking their own initiative to improve their own lot. I've had countless conversations with developers, engineering leaders and product managers around how this insight and openness has enabled better conversations around the need to invest more in technical debt, flow state and other factors impacting developer effectiveness. We've been able to share some internal success stories where teams have made tangible improvements. Then allied to those 'bottom up' outcomes are how this insight enables teams like my own to derive clear product priority signal; find cohorts for user research; define great OKRs and measure the impact of wider initiatives we can support engineers with. The reality here is that these wider initiatives lag and take longer to show up but we are starting to see positive impact. The trick in my experience is to balance how we can help teams with what they are doing today whilst also plotting the course and creating what tomorrow should look like. It's also key that teams feel heard and that centrally driven initiatives feel like they are being done with them not to them. So that bottom up and central out combination is what matters and the sustained participation rate is a side-effect of that being real. Whilst it is true that the way DX have implemented snapshot surveys and the smart nudging via Slack/Teams helps, I think it fundamentally holds true that people will stop participating if they don't feel heard and see positive outcomes from it. That's why I still pay attention to our participation rate and never take it for granted. Aside: the chart shows snapshot participation rates day by day. We launch on a Monday and respondents get a Teams DM. For the subsequent four days people yet to respond get a gentle automated nudge DM. On the following Monday I send a personalised nudge DM that comes directly from my Teams account and this works really well to get a final kick. I also make sure in that nudge to recognise that it captures people who were not at work the previous week and welcome them back :) It's also worth pointing out that people can still respond for some time after we 'close' the snapshot and results are available. We typically see another 1% increase after close. We don't send any emails around calls for participation.