GRO Talks 267

Share to:

Don’t forget, you are awesome. It’s the CRO Roundtable Roundup!

Thanks to Iqbal Ali, Matt Beischel, Slobodan Manić, Shiva Manjunath, Craig Sullivan, and David Swinstead for joining us. Want to get in on the action and talk with other cool CRO people? Then you should probably…

One Sentence Takeaway

I’m a busy person; give me the TL:DR

Running experiments with clear hypotheses, pre-mortems, and risk-aware thresholds helps ensure valuable learning even from inconclusive tests.

The Notes

What we discussed this week

  • Experiment Statistics: Power and Thresholds
    • Underpowered tests often create false confidence in directional results
    • Bayesian vs frequentist approaches to probability thresholds
    • 60% vs 90% significance thresholds and cultural impacts
    • Non-inferiority framing as a valid alternative to strict wins
    • Hypotheses used to permit future decisions despite inconclusive results
    • Expected loss considered better than p-values for de-risking
  • The Importance of Doing Pre-Mortems
    • Importance of writing hypotheses tied to measurable outcomes
    • Pre-mortems encourage thinking through win, loss, or flat scenarios
    • LLMs can be used to generate possible reasons for test failure
    • Predictive bias leads teams to over-assume tests will be winners
    • Pre-mortems help teams prepare iterative next steps
  • Business Metric Trade-Offs
    • Lowering AOV may reduce revenue if not offset by more orders
    • Revenue is a product of both AOV and order volume
    • Some businesses prefer higher AOV, others prefer more customers
    • Trade-offs differ between e-commerce and SaaS models
    • More customers increase long-term remarketing opportunities
  • AI, Data Use, and Compliance
    • Concerns about companies scraping data without consent
    • Past leaks exposed sensitive chats and corporate plans
    • Ambiguity in “make discoverable” wording creates dark pattern risks
    • Companies are intentionally pushing boundaries of data privacy
    • Legal disputes around AI training datasets were discussed
    • Recruiters and employees weigh personal ethics in company choices
  • Dealing with Bug Fixes in Testing Practices 
    • Some large winning tests came from fixing bugs
    • A/B testing bugs highlighted hidden losses and drove improvements
    • Automated testing could prevent costly experiment errors

Hey all you cool GRO People!

GRO Talks Live returns to the winter 2025 Experimentation Elite conference with a heaping helping of in-person roundtable sessions, and we want to see you there!

GRO Talks Live Roundtables
December 9th, 2025
155 Bishopsgate, London, EC2M 3YD

The Quotes

Standout quips from this week

"We make this test valuable where, even though it's inconclusive, it's going to give us something of value."
"Don't torture the data until it comes out the way you want it to."
"You're writing down permission to do whatever past you said he wanted to do."
"Ethics prevents me from, you know, not doing that."
"There's that little bit of dopamine brought into the testing mindset."
"Shocking that this was not about the customers."

Book Club

Relevant reads recommended this week

Empire of AI

Empire of AI

When AI expert and investigative journalist Karen Hao first began covering OpenAI in 2019, she thought they were the good guys. Founded as a nonprofit with safety enshrined as its core mission, the organization was meant, its leader Sam Altman told us, to act as a check against more purely mercantile, and potentially dangerous, forces. What could go wrong?

Off-Topic Sidebars

Experimentation isn’t the only thing we talk about at the CRO Roundtable. There’s often a healthy dose of discussion on shared interests, personal passions, and hobbies.

No off-topic sidebars this week, sorry!

Sidebar Shareables

Amusing sidebar content shared this week

No sidebar shareables this week, sorry!

Share to: