@tibs_at_aiven, @BenGamble7 and I got to go to DevRelCon London this year: Ben gave a talk (“Building internal alignment”) in the You Got This track, and I ran the unconference 2 days, together with Dan Maher from Scaleway.
Together with our friends at Weaviate (
JP Hwang, Zian Hasan, and Marion Nehring) we organized an Open Source Data Infrastructure meetup just across the street from DevRelCon, on September 8. We’ll post the recording shortly!
I also attended the Devrellers meetup on September 7, because “more content is more better”. Emily Omier talked about “How you can give people a reason to join your community”. Emily is a consultant, she hosts The Business of Open Source podcast, a topic she also writes about A Lot, and she’s a frequent contributor to The New Stack.
Resources:
The DevRelCon unconference
An official theme this year, and there were many talks around focus, perf improvement, and proving value, as to be expected. I didn’t get to see many talks, but I’ve summarized the talks I attended below.
For all topics discussed during the unconference, and summaries of the conversations (not available for all): etherpad.mit.edu/p/devrelcon-london
Some returning themes:
- The role of AI in our work, but also the question if it’s our role to control (the use of) AI, is a discussion that was discussed in several slots.
- Also: social media “now that Elon Musk ruined Twitter”. We had the great privilege of having Andy Piper with us, who used to work at Twitter and now does advocacy for Mastodon (although that brings its challenges apparently). Andy explained where the fediverse came from, why certain things are the way they are, and where we might be heading.
My (very lacking) notes from talks (the one I could attend)
Wesley Faulkner (AWS) - The Power of DevRel: A Community of Leaders and Learners
- Community: supportive community of DevRel, find validation outside of the company you work at (now)
- Resilience: disconnect from valuations, remember your core, build resilience
- Perseverance: keep pushing for the right impact…
- … consistently (aka Integrity). The company on your name badge might change, your reputation follows you everywhere.
- Accountability: we all do the right thing always, because we have our own reputations, but you also represent the industry as a whole. Keep eachother honest.
- PAY IT FORWARD: pull other people up with you
We define what devrel is “by doing the right thing, and by always doing the right thing”
We know what isn’t devrel when it’s done wrong. Being uncomfortable through our growing pains is something we can experience together.
Jon Gottfried (MLH) - DevRel Insights - what we learned from 650,000+ developers over 10 years
- 10 years of standardized surveys
- 100k+ hackathon projects
- on the ground feedback
Nearly 3 in 5 MHL community members identify with an underrepresented group in tech.
- Community events can be early indicators of tech trends - What people enjoy using, look for patterns, and support nascent platforms
- The tools in a developer’s utility-belt are extremely sticky, get in early - Cloud platform preference by cohort - Google takes over from AWS as MLH’s official Cloud Partner and the effect is Visible
- Developers have influence over purchasing decisions, even early career devs! 37% MLH alumni introduced technology they learned at a hackathon into Production at their job.
- Giving out good swag created exponential brand impressions: treat your DevRel Brand as a Real Brand!
MLH has an annual mascot and swag refresh, and Jon introduced MLH’s 2024 Season Mascot at DevRelCon: Hanson the Hardware Hippo.
I loved their Jupyter Notebook on The Value of Swag
hackers = 150,000 # annual participants
distribution = 20% # % of attendees that get a shirt
weekly_wear = 36.5% # report wearing theirs weekly
hackers_wearing_weekly = shirts*distribution*weekly_wear
~11,000
daily_sightings = 200 #average number of paths crossed
impressions_per_week = hackers_wearing_weekly*daily_sightings
2.200.200 impressions per week!
Even more DevRel data insights? Dig into an early copy of MLH’s upcoming 2022 Student Developer Skills Report.
Matt Jarvis, Director of DevRel, Randall Degges, Head of DevRel & Community, Snyk - Building a resilient DevRel Org (DevRel in a downturn)
The Tech economy sucks:
- Tech companies generally have highP/E ratios due to higher margins, lower fixed costs, and scalability
- But… since most tech companies have high growth potential, they spend a lot of money to grow more quickly. As a result, they have less free cashflow and more debt
- With rising interest rates, acquiring debt is more expensive, so many tech companies try to cut back spending (layoffs, budget cuts), so they won’t need to borrow money at unfavorable rates.
- With an uncertain economy, large investors are reducing their investment in high-growth tech companies and putting it into safer options (like nomds and more conservative businesses).
How is DevRel impacted? (Google search "DevRel layoffs yields 6.570 results):
- company has limited budget and needs to prioritize other areas (eng, sales, etc)
- Company decides to let go of its most expensive employees (many of whom are in DevRel) to reduce spending.
- Company leadership doesn’t see sufficient ROI for DevRel
Asked OpenAI to analyze all talks at DevRelCon from between 2015 and 2019 that talked about showing value:
- 2015 (London): 20%
- 2016 (London): 27%
- 2016 (SF): 43%
- 2017 (London): 41%
- 2018 (SF): 85%
- 2018 (London): 73%
- 2019 (SF): 73%
- 2019 (London): 78%
We talk about value a lot, yo.
(Typical program) Outputs:
#blog posts#YouTube videos#speaking engagements#community events#technical workshops- …
(Typical program) impacts:
#blog readers, influenced registrations#video views / subscribers, influenced registrations- utm’d link clicks / short URL uses, influenced registrations
#influenced registrations/adoption/ARR/churn reduction#influence adoption/ARR/churn reduction- …
Flexibility: align with company goals
Typical advocate (at Snyk?):
- Staff+ Engineer - with lots of experience and domain expertis
- Influential - is a recognized leader in the space and easily able to act as a trusted advisor to other builders
- Excellent communicator - is a top-tier educator, able to easily convey thoughts, break down problems, and provide technical guidance via both writing and speaking. The type of person people learn a lot from.
- Creative hacker - able to dive into problems with creativity, learn quickly, and build solutions wherever they’re needed.
Able to work across multiple parts of the business to get shit done and remove barriers. - Empathetic user - understands the product and domain in-depth. Can fully empathize with users and understands their needs, wants, and sensibilities. Can help advise product roadmaps, point out adoption issues, and prioritize critical issues.
Regular goals (under good economic conditions):
- Revenue growth
- Market expansion
- Profit maximizatio
- R&D
- Talent acquisition & retention
- Capital expenditures
Downturn goals (under uncertain economic conditions):
- Cost reduction and efficiency
- Cash flow management
- Customer retention
- Diversification and risk mitigation
- Debt management
- Operational flexibility
- Employee well-being
- Preservation of brand reputation
- NPS/SUS/CSAT scores
Be realistic: writing a technical blog = 1 day?
- Research = 2 days
- Writing/editing = 2 days
- Admin= 0.25 days
- Social/GTM= 0.5 days
Writing a technical blog = 4.75 days
Writing a conference talk:
- Writing = 3 days
- Slides/editing = 2 days
- Particing = ??
Preparing a conference talk = 5 days or more
Rule of thumb 45 minutes for 1 minute of public speaking
45*30/60/8= 2.8ish
How much time does an advocate actually have available?
Business as usual activities:
- Meetings = 0.25 days
- Social media = 0.25 days
- Product engagement = 0.25 days
- Technology research = 0.25 days
Average month = 5 weeks
Monthly resource = 5*4 = 20 days
Measure important things. Snyk’s 2020 mission "Get every developer in the world to know about (and hopefully use) Snyk.
- Awareness: How many net-new developers are your DevRel programs reaching
- Acquisition: How many net-new users are your DevRel programs getting to try your product
Measured by Google Analytics
-
Sub-community growth: Python developers, Java ~, JavaScript ~, AppSec engineers, etc
-
What outputs yield impact: How do we build predictability into a model that allows us to always create high-impact outputs?
-
Scale key programs!
-
Ruthlessly automate! And boy, that they do at Snyk.
Measured in “DevRel days”. Generate cost saving based on average Advocate cost per year. Blog reposting saves 60 minutes per blog. With 170 blogs per year thats 170 hours/8=34 days!
DevRel have 20 days per month (remember the model?)
DollarXXXk / 12 / 20 * 34 = 21k per year
Overall program now saves >150 days per year ~ 100k
Richard Millington - Solving community mysteries with data
The best insights arise when you dig as deep as possible into your data.
“Why is the community getting bad satisfaction scores?”
- High response rate, average TTFR, average accepted solution rate
- No major variation by language, location, or month
- Ratings for one product were terrible (“So now I’m just stuck with [product] I can’t use, thanks a lot!”, “You shouldn’t have a community if you can’t help anymore.”)
- Ratings product in customer support were higher than community
Cause: support reps were telling customers with unanswerable questions to post in the community
Lesson: Experience the problem the way the audience does (in this case: shadow support calls)
“Why is the community causing people to cancel?”
- No change in sentiment in the community
- No obvious technology change
- No major changes in cancellations overall
- Suddenly higher visits to 18-month old discussion asking how to cancel
Cause: the link to the cancellation page had been removed from the FAQ and account section.
Lesson:
Spike = Technical change
Gradual shift = Social change
“Why did Kansas suddenly become the most popular source of visitors?”
- Sudden cliff edge in visitors
- All visitors coming from a tiny town with ~180 inhabitants
- Town is geographic centre of the USA
- Analytics tool release notes updated geolocating system
Cause: When the updated analytics tool couldn’t determine IP,it defaulted to the centre of the USA
Lesson: When the data doesn’t make sense, validate the data
Mini mystery 1
- Sudden increase in visitors from India
- Visitors have Chinese names
- But list USA as country in their profiles
Cause: Community was blocked by Chinese firewall. Members used VPN.
Mini mystery 2
- Community had sharp decrease in TTFR
- Many discussions had a negative TTFR
- No obv change in response time visible
Cause: Edited discussions were changing the timestamps of when discussions were posted
Mini mystery 3
- Visitors from Europe plummet
- But post count from European members stable
- No issue with translation tools
Cause: Europeans not accepting GDPR cookie trackers
“Why can’t we sustain event engagement?”
- Annual conference is sold out and feedback is overwhelmingly positive
- Engagement after the event declines and remains significantly lower for several months
- Number of visitors remains consistent, but the number of posts drops
- Decline is primarily amongst event attendees
- Members report connecting with people via other channels
- Easier to use WhatsApp groups
Lesson: Use data to narrow the options, then undertakes qualitative research
Download Richard’s book for free, you won’t regret it: https://qrco.de/belytf