Context

James Daunt has now rescued two iconic bookstore chains: the UK's Waterstones and, more recently, Barnes & Noble in the US. Host Jeff Berman talks with him about how he did it twice with essentially the same playbook, built on a deep distrust of centrally-held data, a flattened, low-blame management culture, and a deliberate refusal to build a marketing department even as BookTok drives real sales through the company's own frontline booksellers.

The Big Idea

A retail turnaround built on curation, not commodity fulfillment, depends on trusting local judgment over centrally-held data, and the discipline that takes, flattening hierarchy, tolerating mistakes without blame, refusing to centrally amplify organic wins, matters more than having a proven playbook.

Daunt's evidence is that his playbook transferred almost perfectly from Waterstones to Barnes & Noble, same customer, same booksellers-as-tribe culture, same underlying mechanics, and yet the turnaround still took just as long the second time, because the real bottleneck was never the plan, it was changing culture at human speed.

Key Insights

1. Gut judgment can beat the data

Daunt describes bookselling as "a qualitative rather than a quantitative judgment" and gives a concrete example: toddlers reading board books want the same few titles read again and again, so a small selection sells enormously well in very little shelf space, while readers over ten want to see everything, so they need maximum space. Many of his predecessors allocated shelf space by raw sales data instead, over-stocking board books and under-stocking young-reader titles, which left toddlers' parents facing an overwhelming, stressful wall of choice while older readers didn't have enough. The lesson: when a decision depends on curation rather than volume, raw sales data can actively mislead you about what the space should look like.

2. Earn the right to centralize data

Daunt calls centrally-held data "this Pandora's box," because whoever holds the data at the center tends to use it to impose uniformity across every location, which he says is "the death knell" for a bookstore even though the same centralization "works in almost every other branch of retail." His prescribed sequence: strip the data out of local decisions first, let that empower stores to build their own judgment about their own customers, and only once that culture is firmly embedded, bring the data back in as a reference point. Reintroduced in that order, he says, data "gets used properly," rather than becoming a tool for control.

3. Culture change doesn't get faster twice

Daunt expected the Barnes & Noble turnaround to be meaningfully faster than Waterstones, since he was applying a directly proven playbook to an almost identical situation, declining sales, disempowered booksellers, a company that had lost confidence. Instead, "the timelines are almost identical." His explanation is that changing culture behaves like stretching a rubber band, "the minute you relax, it just pops back," and that process runs on its own clock regardless of how well you already know where you're headed. Familiarity with the destination doesn't shorten the walk.

4. Skip the blame, just move on

To make a flattened, autonomy-driven culture actually work, Daunt says he deliberately avoids retrospective blame when a reasonable decision turns out badly: "Let's not look in the rear view mirror... Mine is just, okay, was idiotic, move on. What are we doing now?" He argues that this specific practice, not just a stated value of trusting people, is what makes it progressively easier for junior or newer staff to admit mistakes quickly and for senior people to respond without piling on, which keeps the whole team willing to keep making autonomous calls.

5. Zero marketing budget, by design

Barnes & Noble has no marketing department and spends nothing on it, which Daunt frames as an asset, not a gap. BookTok momentum instead comes entirely from young booksellers who are already embedded in local social media, given full latitude to post and experiment as long as they "use common sense, be nice." Books that take off in one store spread "like a rash" to others through bookseller-to-bookseller communication, not a national campaign, which keeps the content feeling authentic in a way a managed campaign couldn't replicate.

6. Selling Kindle was about bookseller fear

During the early-2010s ebook frenzy, Waterstones controversially chose to sell Amazon's Kindle in its own stores. Daunt's real motivation wasn't primarily ebook revenue, it was that his own booksellers had become anxious, believing customers were browsing in-store and then buying on Kindle at home, a pattern he says barely happened in practice. Putting the Kindle "in the middle of it" demystified the device, reassured his staff that the store still had a role as the place of discovery, and turned out to sell an "improbable number" of units as a side benefit.

Mental Models & Frameworks

Catch the wave without gripping it

Berman raises the classic line that "it's easier to catch a wave than create a wave," pointing out that Barnes & Noble is riding BookTok momentum created by its own younger booksellers rather than starting the trend itself. Daunt's refinement goes a step further: even catching the wave doesn't mean touching it. He explicitly declines to build even a lean team to spot booksellers' best organic posts and put paid promotion behind them, arguing that any central involvement, however well-intentioned, risks reintroducing the control that would make the content feel corporate rather than authentic. Use it when a decentralized channel, employee-generated content, grassroots advocacy, is already working: the temptation to formalize and amplify it is usually also the fastest way to break it.

Flatten hierarchy, then hand over responsibility

Daunt's approach to building autonomous teams is sequential, not simultaneous: first actively flatten inherited structures, like the store-manager-to-assistant-manager chain, and only then introduce collective responsibility across the now-flattened team. He notes people often resist this, preferring to be told what to do, but argues the resulting environment is "much more engaging and much more empowering," and, over time, "much more responsive and much more durable" than a hierarchy left in place.

Decision Principles

Principle: Trust frontline judgment over central data

When: the product or service depends on qualitative curation, assortment, presentation, local relevance, rather than commodity fulfillment. Why: centrally-held data tends to get used to impose uniformity, which works fine in most retail categories but actively destroys the local judgment that makes a curated business worth visiting in the first place.

Principle: Don't amplify organic wins centrally

When: a decentralized team or channel, frontline employees, local branches, community members, is already generating attention or results without central involvement. Why: even modest central support, a little budget behind their best post, risks changing how authentic the result feels to the audience that responded to it in the first place, and once that authenticity is gone it's hard to get back.

Principle: Judge decisions forward, not backward

When: a team member's reasonable, good-faith decision turns out badly. Why: assigning blame after the fact makes people risk-averse and reluctant to keep exercising judgment, while quickly naming what went wrong and moving to what happens next keeps the culture safe for continued autonomous decision-making.

Trade-offs & Nuance

Central control speeds up, kills differentiation

Daunt acknowledges centralized data can be "hugely valuable" and that centralizing decision-making "works in almost every other branch of retail." The trade-off is specific to businesses built on differentiated, local curation: the same centralization that makes most retail more efficient and consistent is, in his words, "the death knell" for a bookstore, because it replaces the local judgment that made each store worth visiting with one uniform, centrally-optimized experience.

A playbook transfers, its timeline doesn't

Having already run this exact turnaround once gave Daunt real advantages the second time: confidence in the diagnosis, a working set of moves, and an intuitive sense of which route to take even when it wasn't the fastest-looking one. What it didn't give him was speed, because the actual constraint, rebuilding a culture of trust and initiative among booksellers, doesn't compress just because the leader already knows where it ends up.

Practical Application

Give frontline judgment real veto power

Identify a decision in your organization currently made from a centrally-held dataset even though the person closest to the customer likely has better contextual judgment, such as local assortment, local messaging, or local prioritization, and hand that specific decision to the local team, even where a central system could technically make the call instead.

Adopt a no-blame post-mortem habit

Replace after-the-fact blame assignment following a good-faith decision that didn't work out with a simple two-step response: name plainly that it didn't work, then immediately move to what to do now, without further discussion of who made the call or why. This is the specific mechanic, not just a stated value, that keeps people willing to keep making autonomous decisions.

Resist amplifying organic wins

Before putting budget or central promotion behind content or a program a frontline team member created organically, ask whether doing so would change how the audience perceives its authorship. If a win looks different once headquarters is visibly behind it, let it keep spreading unassisted instead of packaging it centrally.

Reintroduce data only after culture holds

If you've deliberately removed a centrally-held metric or dataset from a local decision to protect frontline judgment, treat that as temporary, not permanent. Once local judgment is consistently strong and embedded as the default, bring the data back as a reference point rather than a mandate, since data used to check judgment functions very differently from data used to replace it.

Questions to Consider

• Where in our organization does a centrally-held dataset currently make a decision that the person closest to the customer could actually judge better, the way board-book shelf space was over-allocated by raw sales data alone at some Barnes & Noble predecessor stores?
• When a team member makes a reasonable, good-faith decision that turns out badly, does our actual practice afterward look like naming what happened and moving on, or does it quietly look like assigning blame?
• Is there a piece of organic momentum right now, a post, a program, a local win, that we're tempted to formalize or fund centrally, and would doing so change how authentic it feels to the people who made it work in the first place?
• If we're relying on a proven playbook from a past turnaround or launch, are we assuming it will also compress this timeline, even though the real bottleneck was never the plan itself, it was culture change running on a human clock?

Bottom Line

James Daunt turned around two different bookstore chains using nearly the same playbook: strip centrally-held data out of decisions that depend on local judgment, let a flattened hierarchy and a no-blame culture make that judgment trustworthy, and then let organic momentum, whether Kindle sales or BookTok, spread on its own instead of being centrally packaged and amplified.

Notable Quotes

"The data can occasionally be an interesting reference point, but it's the judgements you make almost in defiance of the data that will give you an inspiring store." (James Daunt)

"As long as we leave it to the kids with the blue hair, we'll be fine." (James Daunt)