Sh*t happens. And when it happens and you’re responsible for it, how do you respond?
A few weeks ago, Supabase (a backend platform and database for apps) went down.
Their response–both the immediate fix and what they published about it—dramatically increased my confidence and trust in them as a founder with two apps using Supabase.
Whether a software company, a team, or a person, we all fall down here and there. To build reliability and trust in our relationships, we need a certain baseline competence as well as a way of repairing the issue and the broken expectation when breaks happen.
Supabase’s response was a masterclass in repair, so I want to break it down. It doesn’t matter if you run a company or even work with technology: this playbook is applicable to building trust and restoring connection in any relationship where trust matters. After breaking it down, I’ll show you how I applied it with my wife.
Effective, healthy relationships and trust aren’t characterized by perfection, they’re characterized by a cycle of Connection, Break, and Repair.
Our focus here is Repair. Why does this matter?
In interpersonal relationships, when connection is fairly stable and your muscles and mechanisms for repair are in place, breaks in connection need no longer be a source of fear. We can experiment with new ways of being and interacting, have challenging conversations, make important decisions that frustrate others–-and all the while trust that we can cross the chasm of disconnection and find our way back to effective collaboration and connection with each other. We can say goodbye to conflict avoidance and people-pleasing.
Not all Repair involves one party doing something wrong or making a mistake, as Supabase did here, but in cases where we have missed the mark, there’s a “formula” for acknowledging it that I’ve relied on for years, one I see Supabase nailing here. The formula works because it addresses the key questions and concerns the aggrieved party is generally holding:
These 4 steps dramatically increase the likelihood that someone will forgive and trust you sooner.
Supabase’s incident response was written by Paul Copplestone, Supabase CEO and Co-founder. Having the CEO put the postmortem out under his name already conveys a level of gravitas, and it was published the very next day, another sign they’re taking it seriously.
The first three lines establish the first 3 R’s and point to the 4th.
On February 12, 2026, at 21:12 UTC, Supabase experienced a major outage affecting all services in the us-east-2 (Ohio) region. The outage lasted 3 hours and 42 minutes, with full service recovery at 00:54 UTC on February 13.
Recognition ✅
During this period, customers with projects in us-east-2 were unable to access their Postgres databases, Auth, Data APIs, Edge Functions, Storage, Realtime, and any other Supabase service in that region.
Recognition ✅
I am sorry for the impact this caused. We know you depend on Supabase to be reliable. We let you down. This post is a transparent account of what happened, how it happened, and the concrete steps we are taking to make sure it does not happen again.
Recognition ✅
Responsibility ✅
Remorse ✅
But is it believable? Can we trust them again?
Plenty of people will say, “I’m sorry, I feel really bad about that.”
What builds my trust here is the level of transparency and detail, not only about what happened but why it happened.
For example, Copplestone shares:
Why did resolution take 3 hours and 42 minutes? Resolution took an unacceptably long time due to several factors diverting the response team’s attention away from what we ultimately determined to be the root cause…[he goes into the technical details, then shares their management failures, as follows] We did not have representation from the right infrastructure teams at the start of the incident...Had we had the right infrastructure teams involved in the response from the outset, we would have made the connection between the outage and the deployment sooner.
My confidence grows because of the thorough exploration of everything that is within their control, including both technical and management decisions. Resolving the issue immediately is the first step; preventing it from happening again goes deeper still. This is more likely because they explored the system out of which the incident arose and are making edits to that system. That’s where the issue and improvement expands from a technical topic to a management one.
My confidence also grows because they’re not hiding anything.
In any relationship and culture, in any moment of communication, we choose whether to reveal or conceal. When people and companies reveal what’s not flattering to share, even beyond what’s necessary, it fills in gaps in our understanding instead of leaving us wondering. Conversely, if it smells like sh*t and they hang up some roses, I start to wonder what they’re not sharing.
I wasn’t in the affected region, and from what they wrote (“customers in the Ohio region”), I gather that most of their customers weren’t. But instead of sending this only to affected customers and keeping it as private as possible, they publicized it to everybody.
For many people, our reflex when we mess up is to sweep it under the rug because we want to protect our image, but that misses the great opportunity of putting it on the table, building trust, and growing in public.
Ultimately though, trust isn’t about intention; it’s about consistency. Copplestone and Supabase don’t just tell us what they will do, they tell us the plan and what they have already done–two already completed actions (in addition to restoring service), two near-term ones, and five structural changes. And that’s within 24 hours of the incident. The only improvement I would suggest here is an ongoing update on that page with progress to-date.
The 4 R’s formula/framework applies interpersonally. Here’s one way it played out with my wife after we moved countries and I dragged my feet on an important commitment.
She appreciated my candor and level-setting, and I started to restore the integrity of giving my word in our relationship, something I’ve been much more consistent in since then.
It’s only a matter of time until you find yourself in need of a Repair Conversation, and that’s not necessarily a bad thing.
If you’d like more guidance when the time comes, use this custom GPT to walk you through the 4 R’s and see what the conversation might sound like.