Hey, I’m David

🎯 Founder and successful exit of User Conversion; UKs top conversion rate optimisation consultancy
📕 Author of "The Person in Personalisation" 
📍 Founder of Made With Intent: Real-Time Intent Platform
🎆 Big kid. Fan of Disney 

--- 

As ecommerce professionals, we’re living in an era where data and insight have the potential to transform how we connect with our visitors.

We’ve been saying this for years. Using terms like customer-centricity, data-driven and even the dreaded P-word: “personalisation”.

Yet, each year, it feels like we get no closer. We are stuck in an entrenched way of selling that does not align with how people actually buy.

At Made With Intent, we’re on a mission to give you the data and delivery system you need to guide each visitor to purchase. Appropriately. In the most impactful way for them.

If you’re interested in using intent to improve your ecommerce experiences, or just curious about the world of data-driven growth, I’d love to connect.




✅ Former founder building (and successfully exiting) one of the UK’s largest independent conversion optimisation consultancies - User Conversion. 
✅ Global VP of CRO at Brainlabs building world-class conversion product for digital businesses.

David Mannheim

Ben Labay

CEO at Speero | Experimentation & Research Program Leader

CEO at Speero. Recovering Research Scientist. Experimentation Obsessive.

I help SaaS and Ecommerce brands turn data into a competitive advantage. Over the last 8 years, I’ve led research and testing programs for Miro, P&G, ADP, and MongoDB—helping them bridge the gap between complex statistics and better customer experiences.

My mission? To help organizations use experimentation as a superior way to grow.

Areas of Expertise: ◽️ Experimentation Strategy & Scaling ◽️ UX & Usability Research ◽️ Agency & Brand Growth Marketing

Looking for help with your testing program? Drop me a line at ben@speero.com or follow my journey here on LinkedIn.

CXL, and by extension Speero, is an inspired workplace, constantly pushing and questioning services and products and what works and what doesn't. The agency pushes on experimentation and customer experience research for SaaS, Ecommerce, and Lead Gen clients around the world. 

These organizations and teams have business and customer strategies, we test them. Ben works within Speero to scope research and test program strategy for companies like Proctor & Gamble, ADP, Codecademy, MongoDB, and Toast. https://speero.com

I run experimentation programs for companies.  I also… run one for myself. 😄
 
Today’s datapoint:  According to Function Health, my 𝘣𝘪𝘰𝘭𝘰𝘨𝘪𝘤𝘢𝘭 𝘢𝘨𝘦 is 𝟯𝟴.𝟭 — about 𝟲.𝟳 𝘆𝗲𝗮𝗿𝘀 𝘆𝗼𝘂𝗻𝗴𝗲𝗿 than my calendar age of....as of today, 𝟰𝟱. (happy birthday me!)
 
Apparently, my cells didn’t get the memo about birthdays.
 
It got me thinking about how similar my day job and my personal life have become.
 
At work, experimentation looks like:
 
• Forming hypotheses
• Running controlled tests
• Measuring outcomes
• Iterating based on evidence
 
At home, it’s surprisingly similar:
 
• 📊 Fitness tracker to monitor sleep, HRV, and recovery (who's using PAI points?)
• 🩸 Playing with blood glucose monitoring to see how food actually affects 𝘮𝘦
• 🧪 Blood work to track what’s improving, stagnating, or regressing
• 🔁 Small behavior changes → measure → adjust → repeat (𝘵𝘪𝘯𝘺 𝘦𝘹𝘱𝘦𝘳𝘪𝘮𝘦𝘯𝘵𝘴 anyone?)
 
Same mindset. Different system.
 
Whether it’s an enterprise experimentation program or my own health, the lesson is the same:  𝗗𝗮𝘁𝗮 𝗯𝗲𝗮𝘁𝘀 𝗮𝘀𝘀𝘂𝗺𝗽𝘁𝗶𝗼𝗻𝘀. 𝗗𝗶𝗿𝗲𝗰𝘁𝗶𝗼𝗻 𝗯𝗲𝗮𝘁𝘀 𝗽𝗲𝗿𝗳𝗲𝗰𝘁𝗶𝗼𝗻. 𝗔𝗻𝗱 𝘀𝗺𝗮𝗹𝗹 𝗲𝘅𝗽𝗲𝗿𝗶𝗺𝗲𝗻𝘁𝘀 𝗰𝗼𝗺𝗽𝗼𝘂𝗻𝗱.
 
Curious, where else are people running “experiments” in their lives outside of work?

CEO @ Speero | Experimentation for growing SaaS, Ecommerce, Lead Gen

I’ve been thinking a lot about where the experimentation world is headed, and I’ve had this "aha" moment that’s honestly feeling a little bit urgent.

For years, we’ve been comfortable in the world of A/B testing. We know how to test a headline, a button color, or a checkout flow. Even a pricing model or positioning. It’s predictable. But as we start shoving LLMs into every corner of our digital products, from chatbots to dynamic ad copy, we’re realizing that traditional A/B testing just isn't built for this.

The real game-changer isn't just "testing a prompt." It’s this concept of Evals, and if you’re in CRO or product innovation, I feel like you (we) need to be thinking about this 𝘯𝘰𝘸.

The Problem: We’re Testing "Clouds," Not "Bricks"

In the old world, a variant was a static thing—like a brick. You put it there, it stays there. But an LLM is non-deterministic; it’s more like a cloud. It changes every time a user touches it.

You can’t just look at a conversion rate and assume your AI is working. 

Your conversion rate might be up, but what if your chatbot is subtly insulting 5% of your customers or hallucinating fake discounts? (we don't have a lot of this happening yet, but its coming and via chatbots first, have you seen Podium . ai yet? How they pivoted to AI is fascinating.

A standard A/B test won't catch these subtle shifts of the AI until the brand damage is already done.

The Realization: We Need "Agentic" Quality Control. Even synthetic versions.

This is where it gets interesting. To actually control this process, we have to move beyond just "human-in-the-loop" testing (which is way too slow). We need to start thinking about groups of self-evaluating agents.

AI debate clubs on steroids. 

Imagine this: instead of you manually checking if an AI-generated article is good, you set up a team of agents to do it for you. You have one LLM generating the content, another "adversarial" agent trying to find factual errors, and a third "Judge" agent scoring the whole thing against your brand guidelines.
It’s a self-correcting cycle that happens 𝘣𝘦𝘧𝘰𝘳𝘦 the customer ever sees a single word.

Why This is Our Job as experimenters

I truly believe experimentation teams are the only ones equipped to lead this. We’re the ones who already understand hypothesis testing and statistical rigor. 

But we have to shift our mindset from "winning vs. losing" to "quality vs. risk."

We’re essentially becoming the architects of these "eval loops." It’s about setting up the lab (the offline eval) to make sure the AI is safe and accurate, and 𝘵𝘩𝘦𝘯 hitting the real world (the online A/B test) to see if it actually makes money.

If we don't grab the reigns on this process of testing and controlling LLM groups, we’re just going to be cleaning up the mess when these models inevitably go off the rails. 

The workflows and org context are where to start. 

We'll be talking a lot about this at #CircusLondon, May 11-12th. Join us. speero .com / circus

"Product Ops" perspectives are all the rage atm. It's the missing link. It's what most orgs are lacking right now. (We call them XOS here at Speero btw ;)

There are new tools to try to solve this. Tools like Jira's 'Product Discovery' and Airtable's 'Product Central' and even dedicated solutions like Dotwork and Doubleloop. 

We have task management (Jira/Monday), we have wiki's and docs... but what is the tool that allows a team or program to do resource and portfolio management connected to a set of cascaded goals and org-level strategic initiatives? 

At it's simplest, it can look like the pic below here. 

"Testing Learnings," here is a program dash. Do you have a picture like this for your program/team? A perspective across the portfolio of bets that the team is doing? 

This shows a dashboard with several metrics and charts.

> Includes a summary of test results: 9 "Win," 11 "Flat," 5 "Loss," and 0 "Error". 
> displays the monthly relative potential in dollars for discovered wins, discovered saves, and the total. > > Charts showing the breakdown of test results, recommended actions (like abandon, implement, and iterate), and optimization pillars (conversion, frequency, merchandising, and spend)
> Can have charts that link to yearly/quarterly strategic goals as well.... not only the always on stuff but the bigger investments and bets. 

This isn't about a sheet of AB tests that are being worked on. It's a tactical, tangible mechanism that represents and aligns strategy (activities) and outcomes(!) for a team. Boom. It's what helps make a team accountable but also not stuck in an optimization rut.

Recently, a CMO from the supplement industry told us he was working with a performance-based agency. He liked them, and they were making him money. However, the site began to resemble an arcade game. 

He felt that all of the flashing bells and whistles were resulting in cheap conversions, eroding the reputation of the brand as a whole. 

The agency’s singular metric focus backfired when it came to brand loyalty. He needed a more long-term approach to optimization. 

Your customer service can act as a key support optimization factor. You can pull customer data from them, analyze it, fix FUDs, and use it to improve the product, offering, or even the customer service itself.

They provide specific insights into customer struggles, act as a vital feedback loop for website changes (alerting you to increased support tickets), and should even integrate experimentation into their daily work— something that is still surprisingly rare.

Start by integrating support data into your experimentation backlog, automate where it adds value (not noise), and continuously test improvements. 

Efficiency isn't doing less, but learning faster and acting smarter.

#CircusLondon 🎪 AI Workshops, Keynotes, Lightning Talks, Tickets...all coming to the site this week. Stay tuned for the launch and in the meantime register your interest on the LP https://speero.com/circus to get notified when it all drops

𝗪𝗲'𝗿𝗲 𝗱𝗲𝘀𝗶𝗴𝗻𝗶𝗻𝗴 𝘁𝗵𝗲 𝗲𝗻𝘁𝗶𝗿𝗲 𝗮𝗴𝗲𝗻𝗱𝗮 𝘁𝗼 𝗲𝗻𝗴𝗮𝗴𝗲 𝗮𝗻𝗱 𝗰𝗼𝗻𝗻𝗲𝗰𝘁 𝘆𝗼𝘂 𝘄𝗶𝘁𝗵 𝗽𝗲𝗲𝗿𝘀 𝘄𝗵𝗼 𝗮𝗿𝗲 𝘄𝗼𝗿𝗸𝗶𝗻𝗴 𝗼𝗻 𝘁𝗵𝗲 𝗲𝗱𝗴𝗲 𝗼𝗳 𝗘𝘅𝗽𝗲𝗿𝗶𝗺𝗲𝗻𝘁𝗮𝘁𝗶𝗼𝗻 + 𝗔𝗜: 
• 80% in-house/brand-side attendance 
• Workshops, Dinners, and Masterclass events based on core roles
• Thoughtful parties and activation events
• Keynotes by leading in-house practitioners on AI+Experimentation
• 20+ Posters on AI+Experimentation + lightning talks on 10 of them
• 100+ round tables, matching you with people you want to know

Join the show, Circus London May 11-12, an AI conference for experimentation leaders. Participate with a Poster+Lightning Talk on it, covering topics on AI+Experimentation. 

See the doc here, context on page 1, and the poster 'template' on page 2.

The "poster slam" works like this:

1. you have topic and submit an abstract 
2. we get 20 or so from submissions
3. during conference, we print your poster and display it
4. you present a 1 slide, 2-3min talk at the conference on the topic

𝗪𝗛𝗢 𝗪𝗘’𝗥𝗘 𝗟𝗢𝗢𝗞𝗜𝗡𝗚 𝗙𝗢𝗥?

We’re inviting practitioners to submit posters on how AI is being applied inside modern experimentation programs, including:

• AI embedded in experimentation workflows (research, hypothesis generation, QA, analysis, interpretation)
• AI-assisted experiment interpretation & decision systems (judgment models, debate setups, copilots for test outcomes)
• Experimentation operating models augmented by AI (how teams plan, prioritize, run, and learn at scale)
• Research → testing → shipping pipelines where AI reduces friction or increases learning velocity
• AI-driven QA, guardrails, and reliability systems for running trustworthy experiments
• Using AI to scale experimentation (velocity, complexity, orchestration, multi-surface testing)
• Experimentation stacks with AI in the loop (how tools, agents, and automation actually connect)
• Real-world case studies showing how AI supports experimentation in practice - not theory.

If it helps experimentation teams run better tests, learn faster, or make stronger decisions, it belongs here.

𝗪𝗛𝗔𝗧 𝗪𝗘’𝗥𝗘 𝗕𝗨𝗜𝗟𝗗𝗜𝗡𝗚

• 𝗧𝗵𝗶𝘀 𝗶𝘀 𝗻𝗼𝘁 𝗮 𝗽𝗶𝘁𝗰𝗵 𝘀𝗲𝘀𝘀𝗶𝗼𝗻. 
• 𝗧𝗵𝗶𝘀 𝗶𝘀 𝗻𝗼𝘁 𝗮 𝘃𝗲𝗻𝗱𝗼𝗿 𝗲𝘅𝗽𝗼.
• 𝗧𝗵𝗶𝘀 𝗶𝘀 𝗻𝗼𝘁 𝗮 𝘁𝗮𝗹𝗲𝗻𝘁 𝘀𝗵𝗼𝘄.

This is a 𝗰𝘂𝗿𝗮𝘁𝗲𝗱 𝗹𝗲𝗮𝗿𝗻𝗶𝗻𝗴 𝗲𝘅𝗽𝗲𝗿𝗶𝗲𝗻𝗰𝗲 — where experimentation practitioners teach each other how the work is actually done.

Link to submit a poster abstract in comments below!

Circus London Lighting Talk Poster Template (2)

Experimentation COEs have been a big interest of mine for some time, this POV here is something I've been wanting to write down and share for a while now

A year ago, I spent a few months interviewing 19 experimentation leaders across tech companies (even published a paper on it). I asked a bunch of them

Open article: Most Centers of Excellence Become Bureaucratic Graveyards. Here's What the Successful Ones Do Differently. by Ben Labay • 7 min read

Most Centers of Excellence Become Bureaucratic Graveyards. Here's What the Successful Ones Do Differently.

taplio

The Battle

David Mannheim

Ben Labay

Followers

Posts

Engagements

Likes

Performance ScoreBen Labay's LinkedIn Performance

Ready to grow your LinkedIn presence?

taplio

The Battle

David Mannheim

Ben Labay

Followers

Posts

Engagements

Likes

Performance Score.css-goub6s{font-size:var(--chakra-fontSizes-md);font-weight:500;color:var(--chakra-colors-gray-600);display:block;margin-top:var(--chakra-space-1);}Ben Labay's LinkedIn Performance

Ready to grow your LinkedIn presence?

Performance ScoreBen Labay's LinkedIn Performance