Skip to main content
← Back to projects

Case study · Klaviyo

Personalized Send Time

PESTO sends each message when that specific person is most likely to open it. I designed the scheduling UX, manual controls, and analytics that made marketers trust it.

Personalized Send Time
Company
Klaviyo
Focus
Optimization UX · Trust · Analytics
Impact
+6% KAV lift · 27% adoption in 2 weeks

Background

Tuesday at 10am. Everyone. All at once.

Most campaigns send everyone at the same moment. But your customers live on different schedules — and the existing Smart Send Time had gaps: no time windows, an invisible experiment baseline, and results nobody trusted. One large customer (Marley Lilly) churned over this, reporting Attentive's version doubled their revenue.

Design thinking

Build on what people already know.

I tested two entry points. The dropdown won — it plugged into the existing "send type" selector marketers already used, making the new behavior feel like a natural next step.

Toggle — too binary, doesn't communicate a window

Toggle entry point

Dropdown · chosen — familiar pattern, lower cognitive load

Dropdown entry point

Customer selected controls

Make the experiment visible.

The original PST ran a silent A/B test against an invisible baseline. I designed the customer selected controls system — giving marketers visibility into what they're comparing against, without breaking analytics integrity.

Scheduling drawer — define a window, Klaviyo finds the right moment for each person

PST Drawer
Key tension
Too much control reintroduces complexity. Too little and it's a black box. The answer: a sensible global default with per-campaign overrides — and analytics that make the experiment structure explicit, not hidden.

Analytics

Trust is built through proof.

Campaign-level: send time distribution vs. control. Account-level: headline lift across all campaigns. The question marketers ask is "is this working?" — I made sure they could answer it fast.

Campaign analytics — send time distribution vs. control group

PST campaign analytics

Aggregate lift — orders, clicks, opens across all campaigns using PST

PST aggregate lift

Impact

Results

+6%
Average KAV lift per campaign
~0.7%
Projected sitewide KAV lift at scale
27%
Adoption within 2 weeks of GA
~$421k
Estimated annual infrastructure savings
Reflection
Building trust in AI-driven features is as much a design problem as a model problem. When users can see what they're comparing against and understand why a number looks the way it does — that's when they stop second-guessing and start relying on the system.