J&J Snack Foods CMO Lynwood Mallard on the art of keeping the 60-year-old ICEE brand cool

Full analysis

A CPG marketer sat down for a podcast and told you where a mid-market brand budget is going. Retail media at the bottom of the funnel, influencers and user-generated content over display, and a scramble to rewrite websites so AI systems can read them. J&J Snack Foods CMO Lynwood Mallard runs marketing for ICEE, Dippin' Dots, and Super Pretzel. Nothing he said is new. But when a legacy snack CMO says the same thing the whole industry is saying, that is the signal: the mid-market has caught up to the trend, which means the trend is now the baseline.

What's actually being decided: nothing, by us. This is a read on advertiser sentiment. The useful question is whether Mallard's budget moves tell an operator anything they can act on. Easy to undo either way. Nothing sets a deadline. Let me be direct: the direct ad-tech relevance here is low, and I'm not going to pretend otherwise. The interesting thread is one line about AI-readable websites, so most of the council is short.

The Market Analyst Mallard's arc on retail media is the whole story of the category in one guy. He felt "forced to allocate" to retail media networks, the ad businesses run by retailers like Walmart, without understanding why. Now he calls it efficient, measurable, close to the register. That progression, resentment to dependence, is exactly what turned Walmart Connect and its peers into a double-digit-billion line item. When the mid-market stops resisting and starts defending the spend, the growth stops being about persuasion and starts being about renewal. For an operator: the retail media land-grab is past its convincing phase. In plain terms, the hard sell is over because the buyers now sell it to themselves.

The Skeptic Strip the framing and what is here? Self-reported campaign numbers with no outside check: 2,000 locations, 300 earned placements, 2.1 million impressions of user-generated content. An influencer-growth stat sourced to unnamed "third party partners." A claim that storytelling-format web copy feeds AI systems better, pulled from a conference Mallard attended and nothing else. That last one is a working guess with no evidence behind it. AI search behavior depends on the model, the query, and whether the system is even pulling from live web pages. Nobody has a stable answer yet. An operator who reprices content strategy off a snack CMO's conference takeaway is building on sand.

The Customer / End User Here the customer is the CPG advertiser, and Mallard is telling you what he wants to buy. Lower-funnel, traceable, near the point of purchase. Reach-based display buying is out of favor at his shop. That is a demand-side vote against the open web's core product. If a heritage brand with theater and grocery distribution is skewing "a bit more to the bottom of the funnel," the pull toward retail media and away from broad programmatic reach is not a coastal-startup fashion; it reaches the snack aisle and the median buyer. Publishers selling upper-funnel impressions are selling into a headwind that now reaches the snack aisle.

The Operator One thing here you can actually do something with: Mallard says his urgency on making websites AI-readable jumped "like 100%" between two conferences 18 months apart. Set aside whether his fix is right. The demand is real and it is arriving without tooling. There is no clean measurement for whether an AI system quotes your brand correctly, no bid, no standard, no report. A CMO who wants "answer engine optimization" today is buying it on vibes and vendor decks. That gap between a budget that exists and a product that does not is where a measurement or content vendor makes money in the next 18 months.

The CFO Follow the money and this episode is a rounding error made of anecdotes. The Kelce brothers praised Dippin' Dots on a podcast, J&J shipped product to the Eagles locker room, and that is the case study for earned reach. Free is a great CPM until you try to forecast it. You cannot budget a locker-room moment, and Mallard knows it, which is why the actual dollars are moving to retail media where the spend is traceable. The lesson for an operator pricing inventory: buyers will romance earned and organic reach in interviews and pay for the measurable stuff. Sentiment follows the microphone. Budget follows the receipt.

Where they disagree The Market Analyst reads Mallard's retail media conversion as proof the category has won. The Skeptic reads the same interview as unverified self-report and wants nothing built on it. Both are right, because they are answering different questions: sentiment is real and directional, the specific numbers are not bankable. The second split is on the AI-website thread. The Operator sees a budget forming ahead of any product, which is where money gets made. The Skeptic sees a practitioner heuristic from one conference, not settled practice. That tension is the actual opportunity: the demand is early and unserved precisely because the method is not proven.

What it hinges on: whether mid-market CPG demand for making content legible to AI systems shows up as real, paid spend before there is a standard way to measure it. The council leans that the demand is real and the retail media shift is structural, not a fad. What to verify before acting on the AI-content thread: whether any buyer will actually pay for it versus fold it into existing SEO retainers.

Prediction: At least one major retail media network among Walmart Connect, Amazon Ads, and Kroger Precision Marketing will report full-year 2026 ad revenue growth above 20% when 2026 results are published by the end of Q1 2027 earnings season.

Confidence: Medium. Mid-market resistance to retail media has flipped to dependence, but a single quarter of macro softness could dent the print.

Why: Mallard's public arc from feeling "forced to allocate" to calling retail media efficient and measurable is the mid-market advertiser most likely to have resisted, and he has stopped resisting. When the reluctant buyers convert and defend the spend themselves, the category stops relying on new-logo persuasion and grows on renewal plus expansion, which is more durable than a hype cycle. Amazon's ad business and Walmart Connect have printed 20%-plus growth on far larger bases than the snack-brand budgets now flowing in, so the incremental mid-market dollars land on top of an already-fast line. The opposite, growth dropping below 20% across all three, would require the reluctant buyers to reverse the exact conversion Mallard just described, and nothing in the demand picture points that way.

Revisit by 2027-03-31: We're right if at least one of Walmart Connect, Amazon Ads, or Kroger Precision Marketing posts full-year 2026 advertising revenue growth above 20%. We're wrong if all three come in at or below 20%.