From Methodology to Operating System: Why groa° Came to Maybe*
By Polly Barnfield, OBE, CEO of Maybe*
Five years of retention-first work proved the strategy. The constraint was always execution. Here is how that constraint becomes the platform.
Emma Powell had a problem that any consulting founder will recognise. The methodology worked. Brands that adopted Retention-First Growth® consistently moved into Klaviyo's top-decile performance tier. Nine to eleven times higher campaign revenue per recipient than industry average. Repeat purchase rates of 40 to 50% against a 15 to 20% norm. Dormancy held below 30%. The results were repeatable. The case studies were real. The economics were defensible.
The constraint was capacity.
Sustaining that level of performance for a single brand took roughly 70 hours per month. Pulling Klaviyo data, reconciling it against Shopify, layering in third-party loyalty and review tooling, watching for early decay signals across high-value cohorts, intervening within 24 to 48 hours when a Loyalty-orbit customer slipped past their expected repurchase window. Two to three days of manual data work every cycle, before any thinking about strategy could begin. R360 Growth could deliver that for the brands they served directly. The other 97% of the market, the brands who could not afford a specialist agency on retainer, stayed out of reach.
groa° is the answer to that problem. It takes five years of retention-first expertise and embeds it into AI Agents that run continuously, inside the brand's existing stack, governed by the same top-decile guardrails that R360 applies by hand.
That is the build we are working on with Emma and the groa° team.
Why this is a Maybe* problem
Maybe* is an AI Agent platform. We do not build chatbots and we do not build dashboards. We build agents that make decisions, take actions, and operate inside our clients' workflows the way a trusted teammate would. That is the right architecture for what groa° needs, because what groa° needs is not another reporting layer. The book makes this point clearly. Ecommerce already has analytics platforms, execution platforms, and commerce platforms. What it is missing is an operating layer. Something that sits above the stack and decides, moment by moment, what should happen, for whom, and when.
That is an agent problem. Specifically, it is a coordinated multi-agent problem, because retention-first execution is not one decision repeated. It is five orbit-specific decisions running in parallel, each with different signals, different intervention windows, and different economic stakes.
What we are actually building
Without giving away the architecture, the work breaks down into four layers.
The identity and signal layer unifies customer state across Shopify, Klaviyo, and the third-party loyalty and review tooling that brands actually run. This is the substrate on which the Customer Energy Profile™ described in the book is computed in real time, not weekly.
The orbit governance layer is where Retention-First Growth® logic lives. The decision rules for moving customers between Capture, Activation, Value Core, Loyalty, and Reactivation. These are not workflows triggered by static thresholds. They are agent-managed states that respond to behaviour as it unfolds.
The intervention layer translates governance decisions into execution inside Klaviyo. Which content, to which segment, at which moment, through which channel, all within the guardrails Emma has spent five years tuning.
The learning layer closes the loop. Each intervention produces response data. The system uses that data to refine timing windows, escalation thresholds, and content weighting, within defined guardrails. This is the difference between a system that improves and a system that drifts.
What success looks like
For groa°, success is the methodology operating at scale without the 70-hour overhead. For their brands, it is top-decile retention performance without needing a specialist agency on retainer. For the wider market, it is a working answer to the question every Shopify operator is now asking. What does AI actually do in retention, beyond writing more email subject lines?
Over the next three posts in this edition, we will go deeper into how the agents work. Part 2 covers the architecture choices, including why an agent-led model fits retention better than workflow automation. Part 3 walks through a worked example, an autonomous VIP decay intervention, end to end. Part 4 looks at the cognitive partnership between the groa° agents and the human teams who guide them.
Emma's book Retention-First Growth is the place to go for the methodology in her own words. Read it here.
If you have a methodology you want to turn into an agent layer, book a call with the Maybe* team.
