If you listen to brands talk about their customers, you’d think customers move through a structured, step-by-step journey - from awareness to decision to purchase - as if they were following very specific steps. But real people don’t follow steps; they respond to impulses and frustrations in unpredictable ways. That’s where rebundling often goes wrong: companies try to stitch together discrete steps in a ‘customer journey’ instead of designing environments where the desired actions happen naturally. The paradox of customer journey rebundling is that it works best when it doesn’t feel engineered at all. Over-optimizing and improving individual steps while failing to create an experience people want to return to might backfire.
Sephora understands the importance of deep, connected, engaging experiences. The beauty industry is increasingly fragmented - TikTok influencers dictate trends, Amazon undercuts prices, and high-end salons provide personalized service that no retailer can match. Yet, Sephora understands that winning the right to serve the customer doesn’t stop with dominating individual steps across the journey; it requires orchestrating the entire experience. Sephora’s approach to rebundling the customer journey began with a simple observation: that the makeup buying experience was broken. Customers were frustrated, spending hundreds of dollars on makeup foundations that ended up abandoned in bathroom drawers, victims of poor color matching and confusing application techniques. Much like Best Buy, Sephora recognized that helping customers make better choices was more valuable than simply offering more options. Sephora’s solution to this problem involves a simple entry point: a handheld scanner that reads skin tone and assigns a unique Color IQ code. With a few quick questions about the customer’s preferences, the system recommends foundations from multiple brands tailored to the customer’s tone and style. This one decision becomes Sephora’s entry point into the customer journey. By solving a high-friction choice, Sephora builds trust and gathers data to guide future interactions. Over time, it has incorporated more advanced AI capabilities into the original Color IQ and added new entry points into the customer journey. It uses these entry points to create a foothold in the customer journey and rebundle related services, ranging from product tutorials to virtual try-ons. Its digital ecosystem, featuring online communities and makeup tools, helps customers discover, test, and learn at every stage of their beauty journey. What made Sephora’s strategy so effective was its ability to connect the dots across the customer journey. A customer might discover a product on Sephora’s TikTok channel, try it virtually on the Sephora app, read reviews online, get a personalized match in-store, and make the final purchase from her phone, all while staying within the Sephora ecosystem. A young woman walking into Sephora for a foundation match no longer leaves with just a bottle of foundation or makeup matched to her skin tone. She leaves with a Sephora-guided path through the beauty industry. From competition to control I first started looking into Sephora’s impact on beauty retail when a leading European beauty brand invited me to speak to their board, to help them make sense of a problem that was becoming increasingly difficult to ignore. Brand power once dominated the beauty industry, with loyalty tied to names like Estee Lauder or L’Oreal. Brand building was expensive, and few could afford it. Yet, once built, it was a powerful moat. But that was changing. With social media, unknown brands could scale overnight, propelled by influencer hype and viral trends, bypassing traditional gatekeepers. At least, that was the prevailing narrative with which my client began examining the problem. But as they interviewed their target market, they figured that the challenge was more complex. Influencers could help drive demand, but that wasn’t enough. Paradoxically, the path to success for brands was still controlled by a traditional gatekeeper - Sephora. Sephora was no longer just a powerful retailer; it had become a kingmaker. If your product made it onto its shelves, your brand had a future. If not, the road ahead was steep. Power had shifted away from brand loyalty toward owning the customer’s choices, and Sephora was leading that shift, making it nearly impossible for brands to bypass its influence. Sephora follows a three-part playbook to manage this restructuring of power: helping customers make key decisions, locking them into its ecosystem, and earning money from the brands that want access to those customers. First, tools like Color IQ capture customers early by guiding purchase decisions. Second, loyalty programs, in-store exclusives, and digital engagement keep them coming back. And third, brands pay a premium for access to this highly engaged customer base. For all these years, beauty brands had worried about competition from other brands. Now, they were worried about control. Control no longer comes from simply selling products; it comes from owning the customer’s decision and attention. And any company, like Sephora, that meets the customer at their moment of need gains outsize power, turning once-independent brands into dependent players in its ecosystem.
The power shift from beauty brands to beauty retailers can be traced back to understanding the difference between what customers buy and what they actually want. Think about it this way: nobody wakes up wanting foundation or lipstick for its own sake. What they want is beauty, confidence, and self-expression. The product is just a means to that end. Economists have a term for this: beauty brands compete in derived demand, selling products that derive their value from satisfying more fundamental human needs. Beauty brands played this game well. Through powerful branding, they escaped commodity competition by associating their lipsticks and serums with aspirational feelings rather than just functional benefits. L'Oreal wasn't selling moisturizer; it was selling beauty and confidence. But Sephora had introduced a different game. It was selling beauty and confidence, not merely through branding exercises, but by positioning itself as the customer’s trusted advisor. It addressed the direct demand - the primary, emotional, and aspirational needs that drove consumers to seek beauty products - and it did so by creating an immersive ecosystem that guided and engaged the customer. Irrespective of your industry, the conflict between direct demand and derived demand plays a big role in determining control and power today. Companies that own direct demand inevitably dictate terms to those that compete in derived demand. Homebuyers, for instance, dream of houses, not mortgages. Real estate search platforms like Zillow capture direct demand by guiding buyers through their decisions, while banks are often reduced to commodity players, forced into price wars. Sephora understood the importance of controlling direct demand, and with that, gained the right to dominate derived demand. If a customer’s beauty journey began with Sephora, it had the power to decide which brands showed up on its shelves and in its recommendation systems. Owning direct demand is now the strongest position in any ecosystem, and Sephora saw this earlier than most. It built an immersive platform around the customer’s goals. But the brand I was advising had misunderstood Sephora’s strategy. Instead of recognizing the power of serving primary demand and building an ecosystem around the right to serve the customer, they misread Sephora’s success as a channel play. Their instinct, accordingly, was to go direct-to-consumer (D2C) themselves and own their destiny. Inspired by D2C brand Glossier, they believed they could bypass Sephora, build a community on social media, and own the customer relationship directly. A consulting firm had validated this with a 136-page dossier. Glossier, after all, had scaled its business without relying on traditional retail to build a cult-like following. The ‘Glossier dossier’ had now become a bible of sorts internally. By building their own sales channel, they could retain more margin and exert greater control over their brand experience. I told them what they didn’t want to hear: no single brand could outcompete an ecosystem. Sephora already controlled key decision-making moments and combined immersive engagement with the convenience of a one-stop shop. My client, limited to its own products, would spend heavily on customer acquisition, only to offer a narrower, less compelling experience. Instead, I urged them to organize a partner ecosystem by activating their biggest untapped asset: their network of high-end salons and beauty professionals. Instead of trying to claw back control through direct sales, they could more cohesively organize their existing ecosystem of trusted beauty professionals, helping consumers make more confident decisions. The board members nodded politely, but remained unconvinced. It was 2019, and D2C was booming. By 2022, though, the cracks had started to show. D2C didn’t scale well, and customer acquisition costs had skyrocketed. By 2023, Glossier, which had once shunned retail, now desperately needed Sephora. The dream of pure D2C dominance had faded. Within months of getting on Sephora’s shelves, sales were back up for Glossier. A bottle of Glossier You was selling every 47 seconds. Meanwhile, my client had built its direct-to-consumer channel but found itself stuck in an expensive loop, constantly spending on ads just to attract new customers. With the benefit of hindsight, the strategy I proposed back in 2019 now made a lot more sense. When they brought me back in 2023, they were ready for a new approach.
When I returned to work with my client in early 2023, the timing was perfect. ChatGPT had just gone mainstream, and generative AI now offered a practical way to revive the strategy I had proposed years earlier: turning their network of salon professionals into a decision-support infrastructure. By training AI on their proprietary knowledge, brands could now build highly effective AI assistance to support customer decisions. AI assistance, however, carries risk. If not managed carefully, AI assistance can expose brands to reputation damage and even create liability nightmares. Whether it’s beauty products or mortgages, the challenge for customers isn’t a lack of choice, but managing uncertainty about which one’s right for them. An AI assistant needs to provide answers that are trusted, accurate, and legally defensible. My client had two options: either to let customers use the AI directly or to have experts utilize AI to guide customers better. The better strategy, here, was to embed AI assistance within their network of salon professionals and beauty influencers, who already played a key role in guiding consumers but lacked formal assistance. With AI assistance, experts can make more informed recommendations while still exercising their judgment. This choice came with a clear trade-off. When AI supports professionals behind the scenes, it doesn’t need to be perfect - experts can step in, correct errors, and guide the customer. Liability for the final recommendation also shifts from the brand providing AI assistance to the expert using it. This reduces brand risk and allows the AI to improve over time. Once mature, the AI can be used directly with customers. Until then, the human layer ensures safe, trusted decisions. Deploying AI directly to consumers is far riskier. With no expert to catch mistakes, errors go straight to the customer, and the fallout is immediate and, worse still, unintentionally viral. Trust, once lost, is difficult to regain. Deploying AI through intermediaries, such as salon professionals, influencers, and channel partners, is different. Intermediaries have built-in trust and can compensate for AI’s weaknesses by filtering out bad suggestions before they reach the consumer. For my client, the decision came down to balancing two factors: complexity and risk. AI is excellent at handling complex choices - exactly what we needed to build trust in the beauty space. But how safely we could deploy it depended on the risk of getting a recommendation wrong. When decisions are complex but low-risk, like choosing a recipe, AI can assist directly, even if it occasionally makes mistakes. However, for higher-risk decisions, like planning your finances or diagnosing a health issue, accuracy matters a lot more. A wrong answer comes with serious consequences. That’s why industries dealing with high-risk decisions rarely give AI tools directly to consumers. Instead, they integrate AI into expert workflows. A carmaker, for instance, won’t ask drivers to diagnose engine issues using AI, but it will equip mechanics with those tools. The mechanic’s judgment serves as a safety net, catching mistakes before they reach the customer. The brand still benefits from AI’s scale and speed, without risking its reputation. Once companies assess the risk and deploy AI wisely, they can use that foothold to create new control points in the customer journey, or, more powerfully, to wrest control from established incumbents.
This extract is from Sangeet Paul Choudary’s new book, Reshuffle: Who Wins when AI Restacks the Knowledge Economy.
