Notes & best practices from the discussion: thanks to everyone who participated and contributed - I merely collected and summed them up. Please share freely.

Problems & Dilemmas

• How do we put more effort up-front, before the project ever begins?
• There's usually a one-way flow of requests (needs/feature statements) from Marketing / Market Research to Product Development.
• Should I be a product manager or a researcher? What's the balancing act? Is it possible to be both? ("Experience Engineer" as one possible resolution, at Intel ESP because "engineer" provides status equality with "Software Engineer"). What should the relationship between a researcher and a PM be?

Best practices

1. Target your insights to a particular person
2. Mine your client for information about their work, their challenges, their colleagues
3. Make the stories defendable (to other disciplines; go beyond just providing qualitative evidence that is sufficient for our standards)
4. Design trackability of research outputs into the design process; find ways to tie insights into performance metrics as a way to prove ROI of research
5. Find the extra cubicle: developing temporary embeddedness to establish relationships with your client teams. Live with them, become one of them.
6. Identity engineering: Shape your identity carefully based on who your client is (or who you're talking with). "Research" might be too closed a label, and it doesn't help people relate to what you know.
7. Be humble enough to make the idea someone else's (when other people use your words without realising they're yours, that's how you know you've succeeded).
8. Speak like your client, think like your client. (How? See #2 )
9. Know your customer's cadence (their cycles of product development; know when to interject and involve yourself, and how)
10. After presenting, talk about your client's work. How do the insights impact their work? (Take time to socialise the research).
11. Identify the decision makers (both of research budgets as well as product management). Target your insights to them.
12. Understand how the clients define success. (How? This is still a sticky problem that generates vague answers. Probably something you constantly should do throughout a research project)
13. You don't make the company successful, you make the person who hired you successful. Know what they're trying to achieve.
14. Create two kinds of outputs: a. the dog & pony show for impressing the client's clients, and b. actionable/immediate/business-relevant for the client.
15. Find about client's work lives (but be an informed blank slate, so you can quickly ramp up the discussion if they think you should know those things already)
16. When pitching / interviewing, look up the other person on LinkedIn.

   * More discussions at www.journalofbusinessanthropology.com (an Open Access journal).

Also posted on Google+.

Thanks to everyone who participated!