The org began treating the customer's attention as one shared budget. Teams started answering for the customer's total attention together.
Bringing PayPal’s notification overload under control
How I made an ownerless trust problem visible, then set up the council and rules that were still running when I left.
All screens recreated. The flows, decisions, and results are my team’s.
A shared surface with no owner.
Every team at PayPal was spending the same scarce resource, the customer's attention budget, and no one was watching the balance for overdraft.
A research participant told us the new package-tracking alerts weren't helping. She was surprised she'd even opted in, and they vanished into a cloud of competing notifications. That's the failure in one anecdote. More messages don't bring home the point: they just create attention fatigue. The cognitive load tips into noise, and the people most exposed were customers in early life, when trust is still forming.
You can't argue against a problem nobody can see. We had to make the overdraft visible, put it in terms that demanded a structural answer, and find someone to own the balance.
We audited first, then wrote rules, then automated them.
Funding real research took money no one had set aside, and without evidence there was no case for change. I got scrappy, running an informal audit, then pulling together about $100k in dribs and drabs from product leads who'd come to trust me. I brought in an outside firm to run a secret-shopper audit of everything a new customer hears from us, every notification, email, and in-app prompt across onboarding and early life. Once the evidence came from outside and sat on paper, leadership treated it as a risk they had to answer for.
I was told a research project wasn’t an option because there was no budget. I decided to ask a lot of people for spare change and cobble the budget together.
I convened a cross-functional council (product, marketing, legal, content strategy) to analyze the audit findings and author an evaluative framework. They categorized each communication (urgency, regulatory, priority, actionability), which channel it belongs on, and when it should arrive, plus fallbacks for when a channel was blocked, a customer who declined push still received a required regulatory notice another way.
Engineering built a delivery engine that applied the council's framework to every message, holding most in a per-customer queue and releasing them as an orderly agenda, while letting a genuinely urgent message jump ahead. It stayed open to oversight, since the council could sample a single anonymized customer's stream to see how the framework landed, no PII, enough context to judge it.
The output was shared ownership of the customer's attention.
The rules and the delivery engine mattered. The bigger win was a shared understanding that uncoordinated outreach harms the customer, and that someone has to answer for the total, beyond their own channel. I built it to live in process rather than in any one person, and when I moved on, the framework, the intake, and the council were all running.
An evaluative framework (priority, channel, timing) that every message runs through, plus an engine that paces delivery.
In our scored survey research, trust in how PayPal communicated rose nearly 30%.
Keep the evidence fresh.
It isn't clear the rules-based prioritization is still happening. I left a functioning committee, rules, and an engine, but the whole org has to stay vigilant. In lean times it's tempting to bend the rules to goose short-term revenue, and that quietly damages long-term trust.
Building the governance was the right move. Pairing it with a standing external re-audit would have made it more durable; external evidence is what gave it authority in the first place.