All parents of young children on a Sunday night: "It's already 9 PM? Fiiiiiiiiinally...."

Ever since I was a kid I wanted to discuss it with your procurement team

I see 3 reasons that block a team adopting AI solutions. This comes from analysing 5 months of onboarding notes from my team: 1. Inertia. The old way might not be perfect, but it kind of works. So the payoff isn't immediate. They have to invest time now to gain it later. 2. Managers not invested Not joining the Kick-off. Unresponsive to our emails. They might not use the tool. But they have to hold their team accountable. 4. Wanting to go from 1 to 1000 Customers coming from a competitor join the kick-off with campaigns already live. They don't need onboarding, more best practices and knowing what's unique about our tool. But the others? They need a roll-out plan. Training and adoption built up one module at a time. Trying everything at once is like trying yoga in an advanced class. Frustration is the only outcome. Nail these three things and you solve most of your activation problems.

Interviewing is hard, but we make it unnecessarily harder. 5 things that cost you the job: 1️⃣ Weak research Hiring manager, founders, product. Look up the LinkedIn profile, search for podcasts they were invited to. At the minimum, ask Claude to do it for you. Then test the product and ask for access if it's not open to non-clients. 2️⃣ High-level answers Be specific. Most questions are about your experience, so use the STAR method: situation, task, action, result. Write them down today: think about projects you led at work, problems you solved, and explain them using this method. It's your portfolio for your next interview. 3️⃣ Generic motivation For this company, this role, this product. One of these (ideally all) has to be true. Especially in startups, real motivation is more important than deep expertise. We can train for skill, never for will. 4️⃣ Quality of questions Ask questions in every round, but the right ones. Asking HR about target attainment history or the hiring manager about time-off policy is a waste of time and sets the wrong impression. Ask yourself: Who will know this best? Who will care the most that I ask this? And always have at least 2–3 questions for the hiring manager and the founder. It's a huge red flag when you don't. 5️⃣ Not following up For any role, really, but especially for client-facing positions. It's expected and it's part of the job. Always follow up after an interview. Restate your motivation or highlight something that was left unclear. Very few do this, so it's easy to stand out. I've been hiring for 10+ years and lost count of how many interviews I've run. This is why I said no to candidates: 40% for lack of relevant experience 60% for a combination of the reasons above And it's probably 30/70 since most candidates can't give specific answers, leaving me wondering if they have the experience I need. The job market is tough. There are many things we can't control. You have to double down on the things you can.

One of my favorite places in Catalonia: Aigüestortes National Park in the Pyrenees.

If you're reluctant to join an early stage startup due to perceived risk, just this week Hubspot and Zendesk had layoffs. So while the risk is the same, the speed of learning is probably 10x at an early stage.

I get it, you join a startup to grow fast. But growing fast cannot mean asking for a promotion or a different role before you're even finished with the probation period. You were hired to do a job. That means that's what the company needed. Food for thought.

Building a win-back program? Filter for fit before revenue. We pulled a list of all churned accounts. All technically elegible for reactivation. But if you hand that over as a flat list, your team calls whoever's on top because nobody can hold the context in their head. > Why did this one leave? > Were they ever our ICP? > Did we lose them to a competitor, a budget cut, or a champion who quit? So we changed the input. 1/ Claude reads HubSpot fields, call notes and email history. 2/ Then classifies every account by the real churn reason. Product gap. Pricing. Champion left. Company restructured. 3/ Next, filter for fit. Only companies that matched our ICP and churned for a reason we can recover from. Not every logo that ever paid us once. What's left is small enough to work properly and a credible pipeline we can defend. Do you have a win-back program? What's the most common churn reason on your list?

Write down the must-haves for a new role before the first CV lands. I learned it after two rounds of interviews with strong candidates who'd never owned the number the role needed. One had years at two big-name tech companies managing complex partnership portfolios and I nearly talked myself into it. When I asked him to walk me through forecasting renewals, the answer: "I haven't done it, but I've been very close to the commercial team and I ramp up really fast." I believed him and still couldn't hire him. A second or third hire can ramp on the job, because there's documentation and a peer to shadow. The first hire is the documentation. In the very early days you hire for potential (you have no brand and can't pay experienced people) and later you can afford a long ramp. In the messy middle of a rapid growing team you need someone who's already made the mistakes you don't have time to walk them through.

Write down a customer's success criteria in their exact words. "High adoption" is your interpretation. "Our sales director opens the platform on Monday mornings to check the quality of the leads" is theirs and it's the one you can defend at renewal.

Ask every new customer what would have to be true for renewal to be an obvious yes. Then write down exactly what they say. It's the question that stops renewals from turning into arguments about goals you wrote for them.

The real risk with AI agents is giving confident wrong answers. An agent can't flag a renewal risk it never saw or escalate something it doesn't know is an escalation. It answers with whatever is in the customer record. And in most CS teams that record is still a mess.

For every speed number you pay your team on, name the quality number that punishes rushing without quality checks. For example: shipping fast vs number of bugs, activation vs recurrence.

Before an account leaves onboarding, write down what the customer achieved, or why they didn't and what the risk is. Make it visible to the AM before their first call. A status dropdown that says "complete" tells them nothing.

If your VP of CS hands you a single Churn MRR at-risk number, ask what's inside it. Ours was four and a half times higher than our real exposure. A big part of it mixing confirmed churn with deals with medium probability of renewing. Adding the total MRR of all deals with any risk doesn't tell you anything. It just makes the number look scarier than it is and impossible to take decisions. So we built a weighted report. Every deal in the CS pipeline has a renewal status (not started, low/medium/high probability, renewal confirmed, churn confirmed) and each status carries a weight. Churn confirmed gets 100% (of MRR at risk), low probability gets 85%, and something we haven't touched yet gets 10% (to count for surprises). The weighted total is what goes to our CEO. We started running the weighted view in April, on churn, downsell, upsell and reactivation in parallel. Quite happy with it as it has been very accurate. Bottom line, if nobody can tell you the renewal probability behind each deal in your forecast, that number is noise.

Perks of 7am gym: no one feels like talking or being social.