Skip the Strategy Decks. Run Focus Groups. Ship AI That Converts.
Welcome to DX Brief - Retail, where every week, we review industry podcasts and reports to share what’s insightful and what you can do about it.
In today’s issue:
How Your Broken Company Already Knows How to Fix Itself (Retail Transformation Turnarounds)
Sephora - How 100,000 Weekly AI Scans Drive 70% Conversion (And Why That's Just the Start)
How Casey’s Scaled From 300K Emails to 200M Monthly Messages Without Losing its “Hometown Convenience Store” DNA
1. How Your Broken Company Already Knows How to Fix Itself
The Wisdom Of... Show with Simon Bowen, guest Bernie Brookes: The Billion-Dollar Turnaround Blueprint (Sept 28, 2025)
For: Transformation leaders, retail executives, change management teams, turnaround consultants, PE-backed retail operators
The headline: Bernie Brookes has turned around billions of dollars worth of retail businesses across four continents: Woolworths' $5B+ Project Refresh, Myer's transformation, South Africa's largest non-food retailer Edcon (65,000 staff across 17 countries). His secret? The answers are already inside your organization. You just need to know how to extract them. And he's built a systematic framework that works whether you're a solo entrepreneur or running a 23-country operation.
Here's the framework most transformation leaders miss:
Start with brutal truth-telling through focus groups, not consultants. Most turnarounds begin with expensive consultants building PowerPoints. Brookes stays away from them. He runs focus groups with three audiences: staff, customers, and suppliers. Why? Because "companies that are broken, people know why they're broken, what worked, what didn't work."
He literally brings a notepad, takes copious notes, then presents it back to management. "They think I'm a genius. Actually, it's not my plan. All I've done is stolen what people who've been in the business longer than me knew."
Most transformation plans fail because they're built on sanitized data and executive assumptions. If you run an airline and ask customers, they'll tell you the food's cold, seats don't recline, video systems don't work. If you ask staff, they'll tell you rosters are wrong. If you ask suppliers, they'll tell you you're buying the wrong product at the wrong price.
Install an independent scorekeeper BEFORE you do anything else. You need a Program Management Office (PMO) led by someone independent. Not your best executive. An outsider.
Why? Three reasons:
Internal people have relationships that compromise independence,
they carry historical baggage that might be wrong,
they often lack the specific skills needed – Gantt charting, milestone management, measurement systems.
This person sits at your executive table but isn't running a division. They're the referee. They need "numeric skills," proficiency in project management tools, and the ability to be an "honesty broker" between the CEO, board, and transformation teams. Think of them as your transformation air traffic controller—they don't fly the planes, but nothing moves without their clearance.
Most transformations fail because there's no single source of truth on progress. Everyone's got their own version of "how it's going." The independent PMO fixes that. They show up at every transformation meeting, track every milestone, and call BS when plans slip.
Break the master plan into ownership-level plans with three-part milestones. Brookes doesn't believe in one master plan. He believes in "the sum of the parts." You need individual plans for IT, supply chain, people, product, marketing – each owned by a specific leader. But here's the critical piece: every milestone must be measurable (you can track it) and meaningful (it matters).
For example: Don't say "improve customer experience." Say "reduce checkout wait time from 8 minutes to 3 minutes by June 15, measured by POS timestamp data, impacting customer NPS and cart abandonment rate."
What to do about this:
→ Run diagnostic focus groups before anything else. Schedule 3-5 focus groups each with staff (multiple levels), customers (multiple segments), and key suppliers over the next 30 days.
→ Hire or assign an independent PMO lead. This person reports to the CEO but sits outside the operational org chart.
→ Audit your current milestones against the three-part test. Pull up your current transformation roadmap. For every milestone, ask: Is it measurable with actual data? Does it impact capital use, end-game return, or customers?
2. Sephora - How 100,000 Weekly AI Scans Drive 70% Conversion (And Why That's Just the Start)
Fast Company Innovation Festival, Retail Reimagined panel: What’s Next for Sephora’s Boldest Bet (Sept 16, 2025)
For: Retail transformation leaders, omnichannel strategists, store experience designers, specialty retail operators, brand partnership teams
The headline: Sephora is redesigning every single one of its 600+ North American stores over the next few years. But that's actually the least interesting thing CEO Artemis Patrick is doing. She's turned beauty retail into a cultural movement by sponsoring women's sports leagues, producing Hulu documentaries, hosting 8,000-person beauty festivals (during hurricanes), and building a creator platform that just went live. Oh, and they're doing 100,000 AI-powered skin scans per week with a 70% conversion rate. Here's the blueprint she's using to transform retail from transactions to cultural force.
Here's the framework most retailers miss:
Start with belonging, not fitting in – it changes everything else. Patrick leads with a principle her daughter's high school principal That distinction drives every decision at Sephora. When the brand arrived in the US in the late 90s, nobody wanted to work with them. Big brands worked with department stores. So Sephora partnered with indie brands nobody else would touch: Stila, Too Faced, Tarte, Urban Decay. Brands that "did not have an outlet to tell their beautiful stories."
That outsider DNA stuck. Even as a beauty giant, Sephora still thinks of itself as "the little guy." Patrick's team sees their role as giving voice to founders who "believe passionately and deeply that the world needs their product."
When you position yourself as the platform for belonging (not conformity), you make different decisions about partnerships, store design, product curation, and technology.
Most retailers optimize for "best sellers" and "category killers." Sephora optimizes for "whose story needs to be told." That's why they sponsor women's basketball leagues, not just beauty brands. That's why they produce documentaries about musicians and beauty expression. They're not selling products – they're creating cultural space where people belong.
Partner early, integrate deeply – don't just slap logos on things. When Sephora entered women's sports, they didn't become "just another sponsor." They became founding partners. With Unrivaled (the 3-on-3 women's basketball league), they got in during the inaugural season. With the Toronto Tempo (first WNBA team in Canada), they're a founding partner launching next spring. With the Golden State Valkyries, they created an entire content series about players' backgrounds.
The strategic They build glam rooms at practices, tell player stories beyond the game, and create content series that inspire the next generation of female athletes. They're not buying visibility – they're building authentic relationships.
This matters for retail because most brand partnerships are transactional: pay for placement, get exposure, measure ROI in impressions. Sephora measures ROI in cultural impact and authentic storytelling.
When you're willing to partner early (risk capital on an unproven league) and go deep (glam rooms, content creation), you become part of the story rather than an ad in the story.
Use AI to enhance humans, never replace them – then close the loop with data. Sephora's AI-powered Color IQ skin scan is doing 100,000 scans per week with a 70% conversion rate.
But here's what makes this actually work: The technology doesn't replace the beauty advisor – it enhances them. The scan identifies your skin shade and type, then uses AI to recommend products based on what similar customers liked. But it also asks questions: Summer or winter? Light or heavy coverage? The beauty advisor uses the scan data to have a better conversation, not to eliminate the conversation.
Then Sephora closes the loop: They track which samples you took, what you bought, what's on sale that matches your profile. "It's a very personalized experience and we leverage all that data to make the experience really seamless."
Most retailers think AI will replace labor cost. Sephora thinks AI will increase conversion rate by empowering advisors with better data. Those are fundamentally different strategies: one cuts costs and degrades experience. The other increases revenue and enhances experience.
What to do about this:
→ Audit your brand positioning through the "belonging vs. fitting in" lens. Ask: Are we trying to serve the broadest possible customer (fitting in) or are we creating space where specific communities feel they belong?
→ Test AI that enhances advisors, not replaces them. Don't deploy AI to cut labor. Deploy AI to make your existing team more effective. Pilot one AI tool (product recommendation engine, virtual try-on, skin analysis) that gives frontline staff better data to have better conversations.
3. How Casey’s Scaled From 300K Emails to 200M Monthly Messages Without Losing its “Hometown Convenience Store” DNA
Customerland TV, Season 3 - Episode 41: If AI agents run retail, who keeps thinking? (Sept 30, 2025)
For: Digital transformation leaders, CDO/CMOs in traditional retail, grocery/convenience operators, retail technology strategists
The headline: Art Sebastian took Casey's General Store – a 50-year-old Midwestern convenience chain – from sending 300,000 emails per month to 200 million messages per month (emails, SMS, push) reaching 7.5 million subscribers. In five years, they scaled digital engagement by 666x. But the real story is how they did it without losing the "hometown convenience store" DNA that made them successful. Here's the operational blueprint.
Talk to customers first, not dashboards. Sebastian's first move wasn't a technology audit. It was customer conversations. "We stepped out and had real conversations with our customers to get better insight into how they view us and what they need from us moving forward." This happened in his first couple of weeks.
Then he audited the technology focused on customer data: "We had a lot of data that no one used... fragmented in multiple systems... we didn't have a common view of our customer."
Sound familiar? Step one isn't buying more tools – it's consolidating, deduplicating, and cleansing what you have so you can actually see your customer.
Build the team using "two in the box." Casey's had no real marketing team – just some folks doing advertising within merchandising. Sebastian created a "Digital Experience Group" from scratch but used a brilliant staffing model: Staff roles initially with external consultants, simultaneously recruit permanent hires, run them together until the internal person is up to speed, then the consultant steps away.
This took two years to execute fully. The You don't wait 6-12 months to recruit, and you don't risk bad hires learning alone. The consultant trains their replacement.
Gen Z wants personalization, not friendship. Gen Z has That means: Don't show me stuff I don't buy. DO show me things you know I buy regularly. Call me by name. Prepopulate my frequently-shopped stores. Let me control email frequency.
Prior generations saw brands as friends. Gen Z sees through that: "You're just trying to sell me shoes. As long as we can get past that, great. But don't tell me you're my friend." This changes decades of brand strategy.
AI is an era, not a transformation. Sebastian sees three
Predictive AI (historical data predicting forward),
Generative AI (creating new content), and
Agentic AI (agents coordinating with other agents).
Most retailers are "still trying to figure out what it is," so they rely on vendors: "My loyalty vendor, my ecom vendor, bring me the AI."
Looking ahead: "The most progressive retailers will have AI agents talking to other AI agents to simplify the experience. More capability with less people behind the scenes."
The warning: IT shifts from vendor management to integration management. And I have a fear we will see deprecation in critical thinking skills – that we over-rely on AI and begin to lose skill. We must not surrender all of it to technology."
What to do:
→ Run 10-15 customer listening sessions before your next technology decision. Extract themes first.
→ Audit where customer data lives and who actually uses it. Consolidate what nobody touches.
→ Pilot "two in the box" for your next critical hire—consultant trains permanent employee.
→ If Gen Z is 15%+ of customers, audit consent management. Can they control frequency and preferences?
→ Inventory AI use cases by phase (Predictive, Generative, Agentic) and map what problem each solves.
→ Identify decisions that must remain human-owned even if AI could automate them.
Disclaimer
This newsletter is for informational purposes only and summarizes public sources and podcast discussions at a high level. It is not legal, financial, tax, security, or implementation advice, and it does not endorse any product, vendor, or approach. Retail environments, laws, and technologies change quickly; details may be incomplete or out of date. Always validate requirements, security, data protection, labor, and accessibility implications for your organization, and consult qualified advisors before making decisions or changes. All trademarks and brands are the property of their respective owners.
