Give people something to respond to

Meeting someone new is easier when there is something to talk about. An unfinished project can be enough.

You can show what you are trying to build and where you are stuck. That gives someone a way to help, question an assumption, or show you a different approach.

Ask a question someone can answer

"I'm trying to make this work offline. Have you solved that before?" is easier to respond to than "What do you think of my idea?"

Showing a screen and asking someone to complete a task is better still. You learn what they understand without explaining every step for them.

For qualitative research, the GOV.UK Service Manual usually recommends 4 to 8 participants per round for methods such as interviews and usability testing. It recommends more rounds when needed, and much larger samples for surveys or A/B tests. A few conversations can uncover a problem; they cannot establish how common it is across a whole market.

You do not need to have every answer

Research by Brooks, Gino and Schweitzer found that asking for advice can improve perceptions of competence, especially when the task is difficult and the person asked has relevant expertise. Asking a specific question can show that you recognize what you still need to learn.

A broader study of group performance involved 699 people working in groups of two to five. Collective intelligence was associated with social sensitivity and a more equal distribution of conversational turns. That is an association, not a recipe for every team. It is a reason to pay attention to who gets heard.

Leave enough room for the other person to explain. Their question may be more useful than your pitch.

Make the first conversation small

In experiments with train and bus commuters, Epley and Schroeder found that talking with a stranger was more positive than participants expected. The setting was commuting, not a tech event, but the mismatch between expectation and experience is familiar.

My suggestion for a first event is simple: bring one thing you are curious about and talk to one person. That is a starting point, not a research-backed target. You can leave early and still have learned something.

Follow up with the work

If someone helps, send the update. If you promised a repository, share it. Tell them what you tried and what happened.

For an open-source project, a readable README and a clear contribution path let that conversation continue without you being there to explain everything. The Open Source Guides are a useful place to start.

The next room is on our calendar. Bring something people can respond to.

Read more

• GOV.UK, Plan user research for your service. Choosing methods, participant numbers and research rounds.
• GOV.UK, Using moderated usability testing. How to learn from a task without leading the participant.
• Brooks, Gino and Schweitzer, Smart People Ask for (My) Advice, 2015. Research on advice-seeking and perceived competence.
• Epley and Schroeder, Mistakenly Seeking Solitude, 2014. Experiments on the expectations and experience of talking to strangers.
• Woolley and colleagues, Evidence for a collective intelligence factor, 2010. A study of how groups perform across different tasks.
• GitHub Open Source Guides, Starting an open source project. Helping people understand your work and find a way to contribute.