The Relationship Design Conversation
Design how you'll work together before you need to — and assume your founder is terrified.
- Difficulty
- Moderate
- Time to result
- ~weeks to results
- Steps
- 6
- Confidence
- 90%
Berger's prescription for the first-PM (and first-20-employees) problem: instead of assuming you'll just come in and do your thing the way you always have, run an explicit conversation with the founder to design how you'll work together — expectations, what success looks like, what each of you actually wants. It pairs with a decoding lens: if your founder's behavior is confusing, assume they are terrified all the time, and see if their behavior suddenly makes sense.
Origin
Distilled from Berger's own failure as Slack's first PM (he came in overconfident, never asked what success looked like, and was fired three times in a year) and from years coaching both founders and the executives who report to them.
Core principles
- 01If you don't have a healthy relationship with the person you're working for, you will not get anything meaningful done.
- 02The job description was written before they met a real human being with real strengths and gaps. The real role must be co-designed.
- 03Early-stage startups have almost no process or structure — if you want clarity on success, you have to create it yourself.
- 04Founders operate under a permanent sense of high stakes. Assume terror, not malice or incompetence.
- 05This is a design conversation, not performance management. Framing it as 'you're already doing this poorly' kills it.
How to run it
- 1
Drop the naive assumption
Refuse the default posture of 'I'll just come in and do my thing the way I've always done it, it'll be fine.' Replace it with honesty: I actually don't know how I'll operate at this company; I have range, and I don't yet know what this company needs or what my place is among these people.
- 2
Hold the conversation early, ideally at hire
Run an explicit relationship design conversation with the founder/CEO in the first weeks, not after things go wrong. Berger recommends founders do this with each of their first 10-20 employees as standard practice.
Pro tip Frame it forward-looking: what's the best way to work together given who we have now, what we want, and what we need? Let's figure it out.
- 3
Get concrete on expectations and success
Cover what success looks like in this role, what the culture actually is, and what the founder's expectations of you are. Establish metrics for success and, if none exist, build them yourself.
Watch out Berger skipped exactly this step at Slack — instead of getting clear on success, culture, and expectations, he made his best guess and went for it.
- 4
Apply the terrified-founder lens
When the founder behaves confusingly, run the assumption that they are terrified all the time — everything is on their shoulders — and check whether the behavior now makes sense. Work with the state they're actually in rather than wishing they were different.
Pro tip Do this rather than either taking the behavior personally or pretending it's something other than what it is.
- 5
Name your own wants inside the relationship
Say the basic things out loud: I want to do a good job, and I want a real professional partnership with you. Obvious wants left unspoken are still unspoken wants — and you'll behave as if they don't exist.
- 6
Re-run it as things change
Treat the design as iterative — collaboratively figuring out an effective way to work together is ongoing, not a one-time onboarding artifact.
In the wild
Joining Slack in 2014 as its first PM, Berger came in believing he was a pro — ex-founder, iconic products, I know how to do this. He never asked what success looked like, what the culture was, or what was expected of him. But the core team had been through the Glitch-to-Slack pivot and layoffs together and wanted someone humble who would earn their trust.
→ Overconfidence collided with the actual expectation. Six weeks in, he came back from a bachelor trip to an email saying he was fired. He lasted a year and three firings in total.
Berger's standing advice to anyone reporting into a founder: when the founder's behavior is confusing, assume they're terrified all the time and re-read the behavior through that lens.
→ The behavior usually becomes legible — everything is on the founder's shoulders and the stakes feel existential to them — which lets you work within their real state rather than fighting the version of them you wish existed.
Common mistakes
Coming in with overconfidence instead of questions
Prior success ('I've been a founder, I've worked on iconic products') convinces you that you can skip the discovery. But what the company wants may be the exact opposite of what you're bringing, and you won't find out until you're fired.
Framing it as performance management
If the conversation reads as 'you're already doing this poorly, you're in big trouble', the other person defends rather than designs. It has to be a genuine joint figuring-out.
Waiting for the startup to give you structure
Early-stage companies have little process. If you want defined expectations and success metrics and you don't create them, they simply won't exist — and you'll be judged against invisible ones.
Is it for you?
Best for
First PMs, early execs, and anyone joining as one of a startup's first 10-20 employees — and founders making those hires
Not ideal for
Large, mature companies where the role, ladder, and success criteria are already well-defined by existing process
From the transcript
“your founder is behaving, imagine that they're terrified all the time. Right? And see if that makes their behavior more clear.”
“And it's actually something that I I recommend to a lot of founders for, you know, maybe their first, you know, 10 or 15 or…”
“instead of getting really clear on what does success look like here or what's your culture or like, you know, what are your expectations of…”
From the episode
Why not asking for what you want is holding you back
Kenneth Berger (exec coach, first PM at Slack)