Start with the outcome

Start with the outcome, not the format

It's natural to come to Atomic thinking about messages. Should this be a card? What should it say? How should it look?

Often, that's not the best starting point.

Atomic is most powerful when you start with what you want your customer to do – and work backwards from there. A card is just the mechanism. The outcome is the point.

Before you build anything in Atomic, answer these three questions:

The three questions

1. What outcome are you driving?

Be specific. "Improve engagement" isn't an outcome. "Get customers to activate their roaming plan before they travel" is.

Your outcome defines everything else – what the card needs to say, what action it needs to prompt, and how you'll know it worked.

2. How urgent is it?

Urgency determines how and where you surface the card.

• Low urgency – a card sitting in a feed is fine, the customer will get to it eventually
• Medium urgency – the card should be prominent, maybe appear on a relevant screen
• High urgency – the customer needs to see this now, interruption may be warranted

Getting this wrong in either direction has real consequences. Over-interrupting erodes trust. Under-surfacing means your message never lands.

3. What happens if they ignore it?

Every use case needs a fallback strategy. If the customer dismisses the card, or never opens the app:

• Does a push notification follow?
• Does the card resurface after a set time?
• Is there an SMS or email fallback?
• Does someone on your team get notified?

Atomic supports all of these – but only if you've planned for them.

Why this matters

When these questions go unanswered, cards tend to feel generic, get dismissed, and don't move metrics. When teams work through them, the result feels timely, relevant, and useful – because it was designed around a real moment in the customer's life, not around what was easy to send.

A note on interruption volume

Not every use case justifies interrupting your customer. Atomic gives you the tools to be as subtle or as prominent as the moment requires. Part of good Atomic design is having an explicit point of view on interruption. Consider how many cards can a customer reasonably receive in a week before it feels like noise? What's your threshold?

Getting the volume right keeps your customers' trust, as well as making for a better experience.