Everyone slapped “AI-powered” on their homepage in 2023. Parry Malm was building AI that actually worked in 2015.

Before ChatGPT existed, before anyone could spell “generative AI,” Parry built Phrasee – a company that used machine learning to write email subject lines that outperformed humans.

Then he retired. Ate kimchi. Went on walks. And came back this year to build Drumbeat, which does the boring B2B marketing tasks that actually work but nobody wants to do themselves.

Parry spent years sending millions of emails. He’d write subject lines, test them, try to spot patterns in what worked. Some worked. Some didn’t. There was no system to understand why.

When he joined Adestra every single marketer asked the same question: “What should I put in my subject line?”

His answer? “I don’t really know.”

That’s when the light bulb went off. This wasn’t just his problem. It was everyone’s problem. And there was no good solution.

Specific words trigger specific responses in the human brain. When you have enough examples of words and phrases at scale, you can model human behavior. You can predict how people respond to certain language.

That’s what Phrasee did. It built generative language models brand by brand, output human-sounding subject lines, ran tests, and created optimized models that predicted language performance before it was sent.

Here’s the thing most people miss: language IS data. It’s just not labeled that way in our brains.

We think of data as numbers and metrics. But words are data points too. Each word choice has a response attached to it. String enough words together at scale, and you’ve got a dataset you can model.

The beauty of focusing on language? It was universally broken. Every marketer struggled with it. There was no good heuristic system or mathematical model to guide decisions. People just made educated guesses based on vibes.

And critically – subject lines were the most important part of email marketing that nobody had solved. Everything else in your campaign is worthless if people don’t open the email.

Fast forward to now. AI is everywhere. Every vendor has “AI-powered” something-or-other on their homepage.

But here’s what most people implementing AI get wrong: they start with the solution instead of the problem.

Parry’s advice is simple but brutal: Don’t worry about AI. Worry about the problem you’re trying to solve.

AI is not the starting point. The problem is.

When people talk about AI today, they usually mean large language models. So it’s worth understanding what they actually are.

LLMs work like this: companies scraped the entire internet and built a model that predicts what word is most likely to follow another word.

They’re called stochastic parrots.

The stochastic system follows a normal curve. Most of the time, the answer is within something you’d expect. But every once in a while, the answer is completely bonkers.

This is why you can’t just plug in AI and walk away. You need to understand what it’s actually doing, where it’s likely to break down, and when you need human oversight.

Parry’s practical takeaway is beautifully simple:

When you’re trying to solve a problem, don’t worry about the solution. Think about the problem. Think about it deeply. Take in all the information you can.

Then go for a walk in the forest.

The answer will come to you.

(Or take a shower. That works too.)

Connect with Parry on LinkedIn.

Find out more about Drumbeat at thisisdrumbeat.com.

Most retailers think merchandising is just making products look pretty in a window display. That’s visual merchandising. Retail merchandising is something entirely different – and if you get it wrong, you’re bleeding cash through overstocked products that won’t sell and missed opportunities on products customers actually want.

Retail merchandising isn’t about arranging products to look nice. It’s about getting five things right:

Get any of these wrong and you’re not making the profit you should be. Launch a lightweight summer range in winter? You’ve failed. Stock gorgeous tops but forget to buy enough bottoms to match? Your range doesn’t work and customers won’t buy.

Everyone says “know your customer” but most people stop at demographics. Anna’s teams used to have a board in the room when picking products – with specific details about the customer: her name, age, spending habits, how she’d use the product.

They’d literally hold up products and ask: “We love this. Does it fit our customer?” If not, it didn’t get bought.

But here’s the thing – once you’re past startup phase, you’ve got data. And purchases don’t lie. If customers consistently buy a certain type of product, that’s what belongs in your range.

The trick is digging deeper than headline sales figures. You might see “occasion wear made £50k last year” and think it’s working. But dig into the data and you’ll find that revenue came from heavy markdowns, you barely made profit, and the few pieces that did work were versatile items customers could wear multiple ways.

That tells you something useful: keep a small selection of versatile occasion pieces, redirect that budget to products that actually make money.

It’s tempting to just buy more of what sells well. But Anna’s blunt about this: “Wear out bestsellers, they become boring, you’ve saturated your customer base, and then you’re going to start having to mark them down because you’ve sold it to everybody.”

You need variety. New products. Things that stretch your range. But – and this is critical – you also need depth in your bestsellers. Stock out of a bestseller and you’ve lost that sale forever. That customer isn’t coming back tomorrow to check if you’ve restocked.

The balance? Anna uses an 80/20 rule. 80% proven sellers with depth of stock, 20% new products to test and keep things fresh. This isn’t a rigid formula – adjust based on your business – but it’s a framework that prevents both boring your customers and taking wild punts on unproven products.

The shift to fast fashion fundamentally changed customer expectations. Products arrive faster, trends move quicker, and customers expect newness constantly.

For online retailers, this means you can’t just plan big seasonal drops anymore. You need regular product introductions to keep customers coming back. And you need to move faster when something isn’t working.

The advantage of online? You can test products quickly, see what’s working in days not weeks, and react fast. The disadvantage? Your competitors can do the same thing.

Looking at data is useless if you don’t act on it. Anna’s seen too many businesses that review numbers, have meetings about numbers, then do precisely nothing about what the numbers are screaming at them.

The key questions to ask your data:

But here’s where most people mess up: they don’t have their data systems set up properly. They’re spending hours drudging through spreadsheets instead of seeing exceptions instantly. If your data process isn’t slick, you’re reacting too slowly.

Real-time data matters now more than ever. Anna’s blunt about it: “A day here is like a week before. It’s so critical.”

Everyone gets excited about bestsellers. Nobody wants to deal with the products that aren’t shifting.

But here’s Anna’s key takeaway: stock that’s building up is blocking your cash flow, forcing you into distressed markdowns that damage your brand, and preventing you from spending money on products that would actually sell.

If you’ve got a missed sale because something sold out? Frustrating, but that money’s gone. If you’ve got stock piling up? That’s money tied up that you can’t spend elsewhere, and it’s getting worse every day.

Anna’s specific, practical advice: go look at your worst sellers right now. What stock are you sitting on that’s not moving? Take action today. Be ruthless.

The problem won’t get better if you ignore it. Years of experience prove it never does.

Two things:

First: Look at your data today. If you’ve got a problem, deal with it instantly. Don’t have another meeting about it. Don’t wait until next week. Understand what the data says, decide what action to take, then actually take that action.

Second: Identify your worst sellers. Find products where stock is building up. These are your bottlenecks. They’re killing your cash flow and blocking investment in products that would actually make money. Take action on them today – even if that means aggressive markdowns to clear them out.

Chasing bestsellers is exciting. Dealing with dead stock is boring and frustrating. But fixing your worst sellers will improve your business faster than anything else you do this week.

Connect with Anna on LinkedIn.

Hire Anna to help you at annawhalleyconsulting.com.

Learn more about the world of merchandising at the-merchandising-school.mailerpage.io

Andy Griffiths, Chief Growth Officer at Climb Agency, joins Dan to dissect why most eCommerce brands are obsessing over the wrong metrics and paying what he calls the “Google tax” – targeting people who would buy anyway instead of finding new customers.

The metrics addiction problem – Why CFOs with spreadsheets are setting media budgets, and how the pressure for immediate ROI is pushing marketers toward vanity metrics instead of real business growth.

The Google tax exposed – How much of your budget is wasted on branded search terms targeting people who were already going to buy, and why that 10x ROAS might be costing you actual growth.

Breaking the cycle – The education gap between budget holders and budget spenders, and practical ways to shift conversations from short-term metrics to lifetime value.

Test and learn like bees – Why 80-85% of budget should go on proven tactics, 15-20% on exploration, and how to build proper structure around experimentation instead of ad-hoc “we tried something different last Black Friday” approaches.

AI reality check – Where the real gains are happening (operational efficiency and mass variant testing) versus the hype, and why mid-market agencies are suddenly competing with big holding companies.

“We’re all very used to focusing on that kind of glory stat or something that perhaps seems in the immediate short term super important but doesn’t necessarily give you a full idea of actual net revenue gain for the business.”

“It gets very easy to effectively run around preaching to the converted… A complete religious pursuit to bid as much as they possibly can on brand terms would potentially be at the expense of going more aggressively after non-branded keywords.”

“There has never been more opportunities to test something new that could give you a competitive advantage through early adoption.”

Stop avoiding GA4. Two years on, too many marketers are still using their CRM or Shopify as their “single source of truth” instead of properly implementing Google Analytics 4. The data discrepancies are undermining your ability to test, learn, and break away from CFO pressure for simple ROAS metrics.

Yes, GA4 is harder than Universal Analytics. Yes, it needs more customization. But the attribution capabilities and audience segmentation options make it essential for understanding real customer journeys beyond last-click conversions.

Connect with Andy on LinkedIn.

Learn more about Climb at climbagency.com.

Most retail dashboards are expensive ways to avoid making decisions. Oliver Spark learned this the hard way while scaling The White Company from £6 million to £50 million.

And it’s why he now runs Sweet Analytics, helping retailers focus on the numbers that actually grow businesses.

Oliver’s turning point came from a simple but brutal incentive scheme: grow the business 40% annually or don’t get paid. This forced him to understand something most retailers miss – the relationship between growth and profit isn’t automatic.

“I worked out pretty early on that I could only recruit about 75,000 customers, and I was still losing money per customer. But I understood I made money after nine months.”

That realisation led to what he calls “the magic spreadsheet” – customer metrics that predicted whether the business would hit its targets long before the P&L showed the results.

Here’s where Oliver gets contrarian. While everyone talks about retention being cheaper than acquisition, he argues acquisition is actually more important for growth:

“Depending on what you’re wanting to do, it’s very difficult to grow retention. If you’re a jewelry business, you might get 11% of people to come back in the following year. Even your best customers who absolutely love you, you might get 20%.”

The math is simple: most businesses have between 1.2 and 1.5 orders per customer per year. Moving that from 1.5 to 1.6 rarely moves the needle enough to drive real growth.

“How do you grow a business? You grow a business through acquisition.”

Oliver’s experience at The White Company proved what William Sonoma had already figured out: multi-channel customers spend significantly more.

But here’s the key insight: if your consumer is shopping in a channel, you need to be there. Even if it’s Amazon. Even if you think it “cheapens your brand.”

“If your consumer is there, your job as a retailer is to be there because that’s what consumers want.”

The trick? Ban the word “cannibalisation” from your vocabulary. Focus on total profit at the bottom line, not where each sale comes from.

Oliver doesn’t hate attribution – he hates how people approach it. Most want the easy answer to “how effective is my marketing” without understanding the complexity involved.

His advice: start with the fundamentals before worrying about attribution models.

“The only number you need to concentrate on is: are you acquiring new customers at your allowable cost of CAC? 80% of people I talk to don’t know what their allowable CAC is based on lifetime value.”

The platforms may lie, but they lie consistently. If Meta says your advertising is getting better, it is getting better. If they say it’s getting worse, it is getting worse. That’s often enough to make smart decisions.

Cut through everything else and focus on these metrics:

“Those are the two most important things because those drive your profit. If you know those two numbers in the context of your P&L, you can just spend your time looking at those two numbers.”

Stop obsessing over perfect attribution and dashboard complexity. Instead:

Most importantly: remember that retail growth comes from selling more stuff to more people more often.

Connect with Oliver on LinkedIn.

Learn more about Sweet Analytics at sweetanalytics.com.

Most product descriptions read like technical manuals written by robots for other robots. Dave Harland, founder of Copy or Die and self-proclaimed “Word Man,” explains why your copy isn’t working and what to do about it.

The fundamental mistake: Taking manufacturer specs and dumping them on your product pages. Features tell you what something does. Benefits tell you how it improves your life.

Dave’s rule: Every product description should answer three questions:

If someone can read your headline and think “so what?” – you’ve lost them.

Real example: Dave transformed boring bike light specs (“LED technology, silicone housing”) into engaging copy that led with the benefit (“takes balls to get on a bike at night, but these balls could actually save your life”). The result? A product description people actually wanted to read.

Stop guessing about your audience. The best copy comes from listening to how customers actually talk about your products.

Dave’s research process:

Case study: Bramwell Brown clocks. Instead of “invest in classic design,” Dave wrote “bring the wholesome feelings from childhood back into your life.” The difference? He actually read 150 customer interviews to understand what people really wanted.

Three psychological principles that boost sales:

The McDonald’s lesson: Sometimes the best way to sell a horse is to say “horse for sale.” Don’t get clever when clear will do.

Before publishing any product description, ask yourself: Does this pass the “so what?” test?

If your opening line doesn’t make someone think “that’s for me” or “tell me more,” rewrite it. Lead with the biggest benefit that solves a real problem.

Quick wins:

Dave’s closing advice: Make yourself laugh first. If you find your copy amusing or engaging, you’ll attract customers who think like you – and they’re usually easier to work with.

Follow Dave on LinkedIn.

Sign-up for Dave’s newsletter – The Word.

Join Dave’s Write the Funny course.

Hire Dave’s agency, Copy or Die.

In this episode of Browse Basket Buy, Dan spoke with behavioral science expert Richard Shotton about the psychological principles that drive purchasing decisions.

Richard, author of “The Choice Factory” and “The Illusion of Choice,” shared key insights that can help retailers.

When customers face too many options, simplify decision-making by highlighting “most popular” items.

Frame offers to emphasize what customers miss by not purchasing, rather than what they gain.

Introduce a premium option to make your mid-tier offering seem more reasonable.

Break down costs to their smallest units (e.g., “just £1 per day” rather than “£365 per year”).

Help customers create triggers for repeat purchases by linking your product to specific moments or contexts.

Richard’s insights aren’t just theoretical—they’re backed by rigorous studies and have measurable impacts on customer behavior. For example, one study showed that people rated cookies from a jar containing only two cookies higher than identical cookies from a jar containing ten.

The most useful aspect of the conversation was Richard’s emphasis on testing. Rather than blindly implementing tactics, he recommends starting with principles that are both easy to implement and backed by robust research, then testing to see what works in your specific context.

As Richard notes: “It’s not just the presence of a high-priced item that matters, it’s the order in which you see it.” Even small changes—like ordering your pricing from highest to lowest instead of lowest to highest—can drive significant improvements.

For retailers looking to optimize their conversion strategy without major overhauls, these evidence-based psychological principles offer low-risk, high-reward opportunities to influence buying behavior.

Follow Richard on LinkedIn.

Go and buy Richard’s books.

Check out the Consumer Behavior Lab and listen to The Behavioral Science for Brands Podcast.

Hire Richard’s agency, Astroten.

Browse Basket Buy is a podcast / webinar series (they’re almost the same now anyway) that digs into the three things that actually matter in eCommerce: how customers behave, what the data tells you, and what changes actually work.

Each month, we talk to people who know why shoppers do what they do, how to spot patterns in the numbers, and which changes actually boost sales.

RevLifter is an eCommerce technology company that helps retailers boost performance using intelligent offers.

Using data analysis and automation, we enable personalized offers that drive revenue without sacrificing brand value or margins.

Browse Basket Buy is hosted by Dan Bond, who runs Marketing at RevLifter.