What makes a great Propeller

Our product team turned Propel's five values into 11 behaviours, written by the people who do the work. They are now the bar we hire, promote and hold each other to, and they are published on our careers page.

Values don't tell anyone what to do

Propel has five values: founder's mindset, craftsmanship, client obsession, forthright and outsized impact. We believe in them. But a value is a noun, and a noun doesn't tell a consultant what to do at 4pm on a Tuesday when the brief and the budget disagree.

That gap widens as a company grows. A new starter reads "founder's mindset" and has to guess what it looks like on an engagement. Two managers rate the same person against it and reach opposite conclusions.

The prompt to fix it came from Bending Spoons, the Milan-based company behind Evernote and WeTransfer. They publish a document on what makes a great "Spooner": the behaviours they hire for, promote for and hold each other to. Their values sound a lot like ours. What we wanted to copy was the discipline of turning values into behaviour.

How we ran it

We ran the exercise at our product team's half-yearly off-site, alongside a vision session on where product at Propel is heading by 2030.

If you want to run it with your own team, the format is simple:

1. Each person brings one behaviour that makes someone great at your company.
2. They explain why it matters, in a sentence or two.
3. They give a real example of a time they lived it. No example, no behaviour.
4. People read theirs aloud in order of seniority, most junior first and the team lead last.
5. Group the overlaps, name each behaviour plainly, and write a paragraph on what it looks like in practice.

The reading order matters most. When the most senior person speaks first, everyone else calibrates to them. Reversing it gives the people doing the work the floor, and it showed us how aligned the team already was.

We used AI to synthesise the first draft, and Trent Conway shaped it into the final document. One test kept it honest: if you can't tell whether someone is living a principle, the paragraph is wrong, and we rewrite it.

The 11 principles

The team landed on 11 principles. Each one serves at least one of our values, and each describes something you can see someone do.

• Prioritise for impact (Outsized impact): Outcomes, not outputs. Make the trade-off before accepting a requirement, and say so when a brief would ship but change nothing.
• No permission (Founder's mindset): Don't wait for sign-off. Do what the outcome needs, tell leadership what you did, and own it in full if it goes wrong.
• See a problem, solve a problem (Founder's mindset): "That's not my job" doesn't exist here. Offer help before anyone asks.
• Anticipation at its best (Client obsession): Solve today's problem, then build for the next five. Seeing around the corner is our job, not the client's.
• Leave them better off (Client obsession): Never hand over something the client can't run without us. It is measured after we've gone.
• Pragmatic, not dogmatic (Craftsmanship): Pick the approach that fits the problem, not the playbook. Take the unfamiliar engagement so we don't set in our ways.
• Challenge the default (Forthright): Question why a process exists or a task sits with one role. When a better way wins, update the playbook.
• Drive (Outsized impact): Land the change, not just the deliverable. Plan time for the change to stick, and say so in the proposal.
• Aboard the train (Founder's mindset): Get on early when the ground moves. Try AI now rather than waiting to see how it turns out.
• Sweat the detail (Craftsmanship): Capture requirements fully and read everything the AI drafts, every time.
• Resourcefulness (Founder's mindset): Find a path with a lean team and imperfect information, ship, and note what you'd do with more time.

What leaders should take from it

Let the people doing the work write it

A top-down version of this document would have used much the same words and meant nothing. Because each principle came with a story from someone in the room, the team owns them. The same words handed down from the exec would have been compliance.

Write down where you fall short

Our document says plainly that we used to be good at sweating the detail, and that we are now only partly there. That line does more for credibility than any of the others. A values document that only flatters you is marketing, and your people can tell.

Name the tensions

Principles collide. No permission pulls against sweat the detail. Prioritise for impact pulls against leave them better off when the budget is tight. The behaviour we care about most isn't in the list: a great Propeller says out loud which principle wins today, and why.

Define judgement before you automate it

In the same off-site, our team described a 2030 where AI handles more of the delivery and people own the decisions. That shift is already under way on our client work. When agents do more of the doing, what's left for your people is judgement: what to prioritise, what to challenge, and when the output is wrong.

Most organisations have never written down what good judgement looks like. If yours hasn't, your AI program will scale whatever ambiguity you already have. This is the conversation we have with clients across product strategy, AI and software delivery, and it starts with people, not tools.

Publish it, and accept the risk

The strongest argument against publishing is that a public standard becomes a stick when leadership falls short of it. We think that's the point. If we stop living these, the team can point to the page, and so can every candidate who reads it.

Read them in full

The full principles now sit on our careers page, next to the values they serve. Alongside them are the two things we look for from the first conversation: you already use AI in your own work every day, and you are curious by default.

If you read them and think "finally", we'd like to meet you. We hire ahead of need, so the right time to talk is before a role is advertised.

Read the principles and see open roles →

If you lead a team and want to run this exercise, or work out what good judgement looks like before you scale AI, book a discovery call.