Episode 9: Stop spray-and-pray: grow eCommerce sales and profits in 2026

There’s a catch: blunt discounts dent margin now and brand later. Dan’s take: aim for intelligent offers—offers targeted and optimized to change behavior, not give away profit to shoppers who would have bought anyway.

Who should see an offer—and which offer is right?

Dan’s framework lives on two pillars:

• Targeting: decide who actually needs an offer
• Optimization: test which offer works best for that group

He walked through the signals that build intent: source, first-time vs return visit, funnel stage (browse, cart), products and categories viewed, geo, basket value and age. With those, you can estimate intent across a spectrum:

• High intent: likely to buy—don’t discount
• Low intent: unlikely to buy—don’t waste margin
• Mid intent: on the fence—this is your leverage point

As Dan put it, most sites convert around 2% of traffic. That means a huge block sits in the middle. Aim offers there.

Optimization matters too. Dan shared a real test on email sign-up discounts: 5% vs 10% vs 15%. While 15% pushed the top-line numbers highest, the actual winner was 10% once the team looked deeper—email sign-ups that convert, offer redemptions, AOV, and incremental revenue. He calls it offer sensitivity testing. Bigger isn’t always better.

The hair jokes, Chloe’s Fun Store, and why the internet is weird

Right from the intros, personality cracked through. Dan admitted he appeared in Meta ads and got roasted about his hair online: There were no positive comments. It’s entirely negative—then laughed it off: The internet is a weird place, Chloe.

Cue Chloe’s recurring bit: a pretend shop called “Chloe’s Fun Store” in Cornwall near Truro, open 9 to 5, Monday to Friday. Dan used it to paint a vivid picture of digital haggling at scale—what a shopkeeper could do 1:1, a site needs to do for thousands of people in real time.

These running gags kept the panel human while they unpacked serious tactics.

Demand capture vs demand creation: spend where it’s incremental

Jamie’s most-shared advice starts with a split many founders miss: demand capture vs demand creation.

• Demand capture (think Google, Amazon branded search) scoops up people already looking for you. Protect your brand, but don’t blow the budget on clicks that would convert anyway.
• Demand creation (think Meta) warms up new audiences who don’t yet know you.

Jamie’s warning: don’t let branded search eat your media plan. Add guardrails so spend is actually incremental. Then pair channels on purpose—e.g., Meta to create demand, Google to capture it, Email/SMS to close and retain. Map the sequence, not just the spend.

Fix the leaky bucket before you pour more in

Jamie’s second drumbeat: prioritize retention. If 50% drop off from first to second purchase, fix that before scaling top-of-funnel. Work towards the “magic shop” number when a buyer flips into VIP territory. Acquisition costs keep rising; retention is your compounding edge.

Sandeep nodded to the same truth from a CRO lens: if people are coming and you let them leave without trying to retain or learn, that’s crazy. He pushed brands to experiment—constantly.

Experiment like you mean it: benchmarks out, segments in

Sandeep’s mantra is simple: try things, measure them, roll back if they don’t work. And stop worshipping industry benchmarks. A “fashion” benchmark that lumps Louis Vuitton with a fast-fashion startup is noise. Measure growth against yourself, not against a vague median.

Also, stop averaging everything. Your “site conversion rate” is just an average of very different segments: channel, device, new vs returning, geo, category, campaign. If one segment underperforms, that’s where the work starts.

Chloe jumped in with a story: a brand cut promotional campaigns from 12 to 10 per year. The breathing room let them learn, iterate, and actually grow both sales and profit. Sandeep cheered—and pushed a pet peeve: when teams find a winner and immediately move on to the next project instead of doubling down and iterating into the vein of gold they just struck.

Do CRO and marketing need each other? Yes—if you want scale

Chloe asked a spicy one: are CRO improvements the only way to make all your marketing perform better?

Dan was quick: always be testing across the whole funnel. That’s just good marketing. Jamie agreed, with a twist: CRO can only do so much if your traffic is low quality. You need the right blend and sequence to bring in the best prospects—and then CRO sings.

Sandeep added nuance. Some brand plays (billboards, Super Bowl) are hard to measure or optimize in a classic CRO way. But for email, paid social, and on-site journeys, conversion work makes every dollar smarter.

Chloe tossed another overlooked lever onto the table: payment success. She’s prepping a session on credit card failure rates and said the revenue lift from fixing failures can be huge.

Can sales and profit rise together?

This was the founder question, and the panel gave practical paths to “yes.”

• Dan’s take: manage supply and demand with intent data and product data. Raise prices on hero products if demand proves it. Use targeted discounts to move stuck inventory. Show offers only to the on‑the‑fence buyers. That’s how you widen the sales base without nuking margin.
• Jamie’s channel view: define the purpose of each channel. DTC might be your profit engine with strong retention. Amazon might be your reach and demand-capture engine. Agree on different ROI targets per channel at the leadership table, and stop wasting board hours arguing apples vs oranges.
• Sandeep’s CRO angle: he’s seen all four quadrants—help sales but hurt profit, help profit but hurt sales, hurt both, and lift both. If you fix friction (sizing confidence, returns clarity, findability), you invite bigger baskets from buyers already willing to buy. And if you steer demand to high-margin alternatives where fit is right, profit can climb too.

Audience hot seat: Is the age of conversion rate over?

Regular attendee Edward Scott Finnegan threw a big one: how should brands measure success—and is conversion rate dead?

• Sandeep: Definitely not. Conversion rate is a metric, not a destination. Measure growth over time, not one magic number. Ignore broad benchmarks. Segment your view (email vs paid vs organic, new vs returning) and act where gaps show up.
• Jamie: Keep the hero KPIs—revenue and profit—but manage two ratios:
  
  • TACOS (total advertising cost of sale) = ad spend / total revenue, so you see true marketing efficiency on the whole business, not just last-click revenue.
  • LTV:CAC, especially for high-frequency models. Use at least one year of data. Look at a blended annual view and at monthly cohorts. For fresh cohorts, model predicted LTV early, then validate six months in.
• Dan: Don’t overcomplicate. Retailers are trying to sell more things, to more people, more often. Keep that frame on the wall and let metrics serve it.

What they recommend, name-dropped and practical

• Tools and platforms the panel works with in the wild:
  
  • RevLifter for intelligent offers and on-site promotion targeting
  • Webtrends Optimize for experimentation and personalization
  • Shopify and Amazon as core commerce rails
  • Klaviyo for email/SMS segmentation
  • Amazon Brand Tailored Promotions to target cohorts like high intent, lapsed, or high potential buyers
  • The heart of the episode: Intelligent offers, not blanket discounts

• Testing to run:
  
  • Offer sensitivity: 5% vs 10% vs 15%—but judge on incremental revenue, redemption, and AOV, not just sign-up volume
  • Price elasticity (promotional price testing) on slow movers
  • Segment-specific journeys: channel x device x lifecycle
  • Payment success improvements and card failure fixes

• Strategy moves:
  
  • Set a clear brand stance on discounting—never discount, discount-first, or rules-in-the-middle. Then enforce it.
  • Define channel roles (DTC for margin and loyalty; marketplaces for reach and capture) with distinct ROI targets.
  • Reduce campaign count to make room for learning cycles.

The button test is dead, long live proposition and findability

Sandeep joked they “milked button testing in 2012.” Today, big wins come from:

• Getting the right proposition in front of the right segment (returns, delivery speed, local trust signals, ratings, support)
• Fixing core UX issues like navigation and recommendations, especially for large catalogs
• Providing sizing help and returns clarity in fashion to raise confidence and basket size

The recurring themes the audience loved

• The “leaky bucket” image for why retention must come first
• Chloe’s Fun Store in Truro (open 9–5, Monday to Friday)
• Dan’s hair saga and the reminder that the internet is, indeed, weird
• Be friends—but not best friends—with your product team, so you stay honest and effective
• Test. Learn. Iterate. Then iterate again when you find something that works

What to do next: Focus where behavior actually changes

• Decide your discount stance—and write the rules. Are you “never,” “always,” or “situational with guardrails”? Then target offers to mid‑intent shoppers only.
• Split spend into demand creation and demand capture. Put guardrails on branded search so you don’t pay for what you’d get anyway.
• Fix retention before you scale. Identify the “magic shop” count and build programs to get customers past it.
• Kill the averages. Segment your metrics by channel, device, lifecycle stage, and category. Act on the gaps you find.
• Run offer sensitivity tests, but crown winners on incremental revenue, not surface metrics.
• Tune payments. Even a small lift in card approval and retry logic can drop real money to the bottom line.
• Set channel roles and ROI targets. Stop comparing DTC and marketplaces the same way.
• Keep iterating when you hit something good. Don’t move on because a roadmap said so.

As Dan put it, your work is to sell more things, to more people, more often. The way to do it in 2026 is to change behavior with targeted offers, prove incrementality with tests, and let your channels each do the job they’re best at. Or, in Jamie’s words: bring in the right customers, and CRO gets easy. And from Sandeep: treat people like people—then measure what changes.

‍