The Influence-Graph Fanboy Loop
Run PageRank on your community's network, then approach the central nodes for feedback — not distribution
- Difficulty
- Moderate
- Time to result
- ~months to results
- Steps
- 5
- Confidence
- 92%
How Figma got its first users. Field scraped the Twitter social graph of the design community, ran PageRank to identify the central nodes, then cross-filtered that list against whose WORK he personally admired — and reached out to that intersection asking to buy them coffee and get feedback. The frame is the counterintuitive part: it looks like an influencer growth hack, but it was run as a learning and feedback loop. The evangelism was a byproduct of genuinely wanting their critique.
Origin
Dylan Field's own play, run in 2012-2013 in Figma's earliest days. The network-visualization inspiration came from work he saw with Gephi while interning at LinkedIn; the ranking method is Google's PageRank applied to the Twitter follow graph. Field notes the exact mechanic is no longer reproducible on Twitter/X since the API was shut down.
Core principles
- 01Communities have measurable centrality — the influential nodes can be computed, not guessed.
- 02Centrality alone is the wrong filter; intersect it with people whose work you genuinely admire.
- 03Ask for feedback, not adoption. Adoption follows feedback; the reverse rarely works.
- 04This works best in communities that are good at giving feedback — designers are, most groups aren't.
- 05The ask is small and cheap for them: a coffee, not a commitment.
How to run it
- 1
Build the network graph of your target community
Assemble the follow/connection graph of your target community from whatever public network data you can access, and visualize it. Field used a Twitter scrape and Gephi-style network visualization.
Pro tip The specific platform matters less than the graph. Any community with public connections — GitHub, Substack recommendations, LinkedIn, conference speaker lists — has a computable structure.
Watch out Field's exact route (the Twitter API) is dead. Do not assume the tooling transfers; the method does.
- 2
Run centrality to find the real nodes
Run PageRank (or equivalent centrality) over the graph to find who the community actually orbits — which is not the same as who has the biggest follower count.
Pro tip Field says he has run this on other communities too, out of pure curiosity about social network dynamics. The output is a map of who the community listens to.
- 3
Filter the central list by genuine admiration
Do not take the top N by rank. Look at their actual work and keep only the ones you are personally inspired by. This is the step everyone skips and it is the one that makes the outreach work.
Pro tip Field had literally been tracing Tim Van Damme's icons and using them as test cases in neural network work before ever meeting him — the admiration was pre-existing and specific.
Watch out Approaching high-centrality people you do not actually respect produces a cold, transactional pitch that reads as exactly what it is.
- 4
Make a small, honest ask — coffee and feedback
Reach out with a low-cost, high-sincerity ask: can I buy you a coffee. The purpose you carry into the meeting is to learn from them and get their critique on the product, not to convert them.
Pro tip Field's framing: it started 'more as like a me Fanboy me getting feedback.' The sincerity is the strategy, not a cover for one.
- 5
Convert critique into product, and critics into evangelists
Show them the product, absorb the feedback, and ship against it. Adoption and evangelism come as a consequence of them seeing their own critique reflected in the product.
Pro tip Run deep, uncomfortable user studies with them — Figma ran one with a designer that took hours and needed a bottle of wine, because typing a sentence in Figma was so slow at the time.
Watch out This only works if the community is actually good at giving feedback. Field is explicit that designers are unusually good at it and 'it turns out that not everyone is good at giving feedback.'
In the wild
Field had been tracing Van Damme's icons and using them as test cases in neural network work — a total fanboy. He met him at Dropbox, showed him Figma, and the relationship started from admiration rather than a growth funnel.
→ Van Damme later joined Figma and did the icons for UI3 — a central-node designer converted into a team member, not just a user.
Figma ran an early user research session with a designer from the target list. Text editing was so slow that typing a single sentence took forever; the session ran for hours and required a bottle of wine to finish. He then wrote Figma a very long document of everything he wanted to see.
→ A brutally detailed product critique from a respected community node — exactly the artifact the loop is designed to produce.
Field personally installed and set up Figma for the team, drove home, learned they had an issue (it turned out to be Wi-Fi), and turned the car around to go back and fix it. Shishir Mehrotra did not realize at the time that they were Figma's first customer.
→ Figma's first customer team, and a long-running mentor relationship — earned by absurd, unscalable founder-level service.
Common mistakes
Running it as a pure growth hack
Field notes people always retell this story as 'find the most influential people and convince them to use your product'. That version fails, because the influential people can tell you want their audience. The feedback frame is what got the meetings.
Optimizing for reach instead of admiration
Taking the top-ranked nodes regardless of whether you respect their work produces outreach with nothing genuine to say. The admiration filter is what makes the coffee meeting worth their time.
Assuming your community gives good feedback
Designers are exceptionally good critics. Many communities are not. If your target group can't articulate what's wrong, the loop returns nothing usable and you have burned your best contacts.
Is it for you?
Best for
Pre-launch founders building a tool for a tight-knit, opinionated, publicly-networked professional community (designers, developers, writers) who need brutal early product critique
Not ideal for
Broad consumer products with no coherent community graph, or communities whose members cannot give structured critique
From the transcript
“I thought it'd be interesting to try to like you say look at who the disign network was who the central nodes were which you…”
“so I I kind of constructed this list of like here are the most Central designers in the graph but also then I like looked…”
“the ones I was inspired by I reached out to and said hey can I buy you a coffee and most them are really kind…”
“I think it started honestly more as like a me Fanboy me getting feedback”
“I think it especially works for designers that way because designers are really good at giving feedback like it turns out that not everyone is…”
From the episode
Dylan Field live at Config: Intuition, simplicity, and the future of design