The Mom Store

01Overview

The Mom Store is a Shopify direct to consumer brand serving parents and young families, spanning maternity wear, newborn essentials, kids clothing and curated gifting. My remit is the digital product itself, the storefront as a system rather than a set of pages, across both the website and the shopping app. That means holding the roadmap, deciding what gets built and in what order, and being accountable for the metrics that sit between a first impression and a completed order.

Unusually for a single role, it spans the full arc: how customers discover the store, how they move through it, how they check out, and how the team behind it understands what's happening. Product, growth, search, and analytics aren't separate functions here. They're one continuous surface, and this case study follows that surface end to end.

02Business Challenge

A growing D2C brand runs into the same tension repeatedly: demand and traffic climb faster than the operational machinery underneath them. Discovery, merchandising, checkout, and reporting each start to strain, and it becomes unclear which of them is actually costing the most growth.

• Discovery increasingly happening outside classic search, in answer boxes and AI assistants the brand had no strategy for.
• Friction between arriving on the site and completing checkout that manual review couldn't reliably locate.
• Reporting that described the past accurately but arrived too slowly to change the present.
• A sprawling stack of third party apps, each solving one thing, none of them talking to each other cleanly.
• A roadmap competing for attention against day to day firefighting.

03Discovery

I began by mapping the store as a funnel rather than a catalogue, instrumenting each transition from impression through product view, cart, checkout, and purchase, and marking where the biggest dropoffs actually lived versus where the team assumed they lived. Those two maps rarely match, and the gap between them is usually where the roadmap should start.

04Research

Research ran on two tracks. Quantitatively: behavioural analytics, search query data, and onsite paths to understand where intent formed and where it broke. Qualitatively: reading the actual language customers used, in queries, in support, in reviews, because that language is what both people and answer engines match against.

• Query and keyword landscape across the category and its long tail.
• Customer journey mapping from first touch to repeat purchase.
• Competitive teardown of how rival stores structured discovery and checkout.

05Strategy

The strategy was to treat three usually separate problems as one system: be found (search, everywhere it now happens), convert cleanly (a checkout with the friction engineered out), and see clearly (reporting fast enough to act on). Each reinforces the others: better discovery is wasted on a leaky checkout; a great checkout is wasted if reporting can't tell you it's working.

Everything downstream, the roadmap, the experiments, the automation, was sequenced against that three part thesis.

06Execution

Execution ran as a rolling roadmap of prioritised bets, each shipped, measured, and either kept or reversed. Cross functional by necessity, working across content, design, development, and a stack of specialist vendors to move an item from hypothesis to live change, on both web and app.

• Redesigned the product detail page and shipped a new site theme, with a “Shop by Brand” surface spanning 45 brands.
• Rebuilt the shopping app on Shopify's AppBrew platform, owning the pain-point audit, wireframes, navigation and review rounds through to launch on both stores.
• Stood up an accelerated checkout, plus OTP and social login to cut the friction before the payment step.
• Relaunched the loyalty and rewards programme (tiers and points), migrating it onto a new platform.
• Added product bundles and combos, on-site and in-app search, estimated delivery dates and order tracking, and lifecycle journeys and pop-ups.
• Ran vendor selection end to end, evaluating roughly thirty apps and integrating the ones that earned their place.
• Leadership — led a blended delivery team around a single roadmap: in-house SEO specialists, a content partner, an SEO consultant, and an outsourced development team.

Underneath the feature work sat the operating principles: roadmapping against funnel impact rather than novelty, structured experimentation (one change, one hypothesis, one measurable outcome), and continuous journey optimisation read from live behavioural data. A large part of the role is orchestration, keeping specialists and external partners pointed at the same priorities and unblocked.

07Product Decisions

The hardest part of the role isn't generating ideas. It's deciding which not to build. Prioritisation ran on a simple frame: expected impact on the funnel, weighed against effort and reversibility. Cheap, reversible experiments shipped fast; expensive, one way door changes earned deeper scrutiny first.

08SEO · AEO · GEO

Discovery is no longer one channel. I treat it as three overlapping surfaces, each requiring its own thinking:

• SEO, the classic foundation: technical health, site structure, and content mapped to real search intent, so the store ranks where customers still type queries.
• AEO (Answer Engine Optimisation), structuring content so answer engines and featured surfaces can extract clean, correct answers about products and the brand.
• GEO (Generative Engine Optimisation), making the brand legible to AI assistants that increasingly summarise and recommend before a customer ever reaches a page.

Because discovery now includes AI agents as well as crawlers, I maintain the storefront's machine-readable layer so the same catalogue is legible to a search engine and to an assistant: XML sitemaps and robots.txt for classic crawlers, plus llms.txt and Shopify's agents.md storefront-discovery file for agentic and AI experiences, all kept in sync with clean structured data and schema.

On the technical foundation, that work lifted the site's health score from the mid-60s into the mid-80s, alongside consolidating legacy URLs with 40+ 301 redirects and keeping a near-daily publishing cadence mapped to real search intent. Treating all three engines as one discipline is the point: the same clarity of structure and language that ranks well is what lets a model quote you accurately.

09Analytics & Dashboarding

Measurement is the nervous system of the whole operation. I built the reporting so the numbers that matter, funnel health, search performance, conversion, sit in one place and update fast enough to change a decision this week, not next quarter.

• GA4, Search Console and Semrush wired into a single view of discovery to purchase.
• Dashboards framed around decisions, not vanity metrics.
• Weekly ranking and AI-visibility tracking, so drops are caught in days, not quarters.
• Experiment results read consistently, so “did it work?” has an honest answer.

10AI Integration

AI earns its place on the repetitive middle of the work, the tasks that are necessary but don't need a human's judgement every time. Used well, it widens how much ground a small team can cover: drafting and structuring content at scale, accelerating research and analysis, and supporting the search work across all three engines.

11Automation

Alongside AI, plain automation removes the manual glue between systems, the copy and paste, the weekly export, the report that someone used to assemble by hand. That runs from workflow automation (piloted on n8n) through to keeping inventory in sync across the storefront and the warehouse. Every workflow automated is attention returned to the team for the decisions that actually need them.

12Business Results

The work is ongoing, and it moves the numbers that matter most between a first impression and a completed order:

The headline is efficiency: visitors converting at a materially higher rate, which is what a storefront looks like when discovery, the product page and the checkout are treated as one system rather than three separate projects.

13Lessons Learned

• Discovery and conversion are one problem, optimising either in isolation leaks value at the seam between them.
• The fastest wins came from measurement, not from building. You can't fix a dropoff you can't see.
• AI and automation compound: their real return is the roadmap time they free up, not the tasks they replace.

14Next Steps

Deepening the generative engine work as AI mediated discovery grows; extending experimentation into personalisation; and pushing more of the reporting toward being self updating, so the team spends even less time describing the business and more time changing it.