Impact sizing Experimentation 1hr saved Medium

Most lifecycle teams size experiments after they commit to building them. By then, it's too late to cut the bad ones.

Allison Bryant, Sr. Manager, Lifecycle Marketing at GlossGenius, built a Claude skill that runs the math upfront, and she was kind enough to share what she built and how she uses it. With that, I'll let her take it from here:

The problem: too many experiments, no sizing

" My team has a TON of experiment ideas. We needed a way to quantify which ideas have the best shot at driving the biggest impact, so we could prioritize them.

impact-sizing.skill 
# Impact Sizing Skill

I'll take you through a short structured conversation to gather your experiment parameters, then build a downloadable Excel model you can explore and share.

## Step 1: Primary Metric
"What's the primary metric for this experiment, and exactly how is it calculated — numerator, denominator, time window?"

## Step 2: Audience Segments
"What are the audience segments for this experiment, and how is each one defined?"

Built by Allison Bryant — Sr. Manager, Lifecycle Marketing at GlossGenius

↓ Download skill + Submit a skill

Got a skill? Share it with the community.

Impact sizing in 2 minutes instead of 1 hour

" The Impact Sizing skill takes a campaign brief (or a quick back-and-forth conversation) and builds a fully formula-driven spreadsheet model to predict the conversion impact of a test idea. It handles three model types automatically: single-metric experiments, two-metric funnels, and promotional revenue viability with sensitivity analysis. What would normally take 30-60 minutes of spreadsheet setup (getting the formulas right, color-coding inputs vs. outputs, adding notes, building a sensitivity table) comes out in under two minutes as a clean, ready-to-share .xlsx file.

Evaluating a real onboarding intervention

" I used it to evaluate two potential interventions in upfunnel actions during our onboarding flow (should our first message drive users to download our app or to set up their booking site?) I gave it the audience sizes, baseline conversion rates of the upfunnel actions, and the rate at which users who performed the upfunnel actions subscribed. The model showed me that 5% lift in one metric resulted in 20% more conversions than the other, making my decision easy.

Impact sizing skill output

This is a version of the skill output, comparing two different test ideas in terms of the subscriptions each will drive. You can edit the blue cells if you want to change your assumptions (maybe you think one test can drive 2x the relative lift of the other - you can edit and see which test comes out ahead!)

What you need before you start

" You'll need to know your own baseline conversion rates, audience sizes, etc., to make full use of this skill!

The bottom line

Most teams treat metric definition and audience sizing as a box to check after the experiment is already on the roadmap. Running this skill upfront forces that conversation earlier.

" We can figure out which experiments will help us hit our goals faster.

Reply

Avatar

or to participate

Keep Reading

© 2026 Lifecycle Leaders | The premier community & resource for lifecycle marketing.
Report abusePrivacy policyTerms of use
beehiivPowered by beehiiv