WITHIN CEO Joe Yakuel Says Most Agencies Still Think Like Vendors

Most performance marketing agencies still sell by channel. They pitch Google expertise, then Meta expertise, then a platform-by-platform menu, and leave the client to stitch the results into something that resembles a business. One agency founder built his company to reject that model outright.

---

Joe Yakuel spent his early career on the brand side, at Quidsi before its sale to Amazon and later at The Vitamin Shoppe, watching agencies optimize channels without asking which channels drove the business. That frustration became the founding premise of WITHIN, the performance marketing agency he launched in 2015 and still runs as CEO. Speaking exclusively to ClickZ at Shoptalk Fall in Nashville, Yakuel laid out the thinking behind WITHIN’s model, his read on live shopping and AI agents, and the budget question he thinks every CMO should be asking before 2027 plans lock.

WITHIN is built around the business, not the channel

Yakuel’s complaint with the industry wasn’t a specific channel failure. It was a mindset. “I was always so surprised how much agencies were operating like vendors instead of partners,” he said. Traditional shops showed up with a Google plan and a Meta plan. They rarely asked what actually moved the business: lifetime value, margin by SKU, repeat rate. WITHIN was built to start from the business objective and work backward into channel strategy, rather than the reverse.

Ten years on, Yakuel sees that gap widening rather than closing, since marketing has only grown more fragmented since 2015. He treats the complexity as an opportunity. The faster things change, he argues, the more value a genuinely adaptive agency can bring to a client’s table.

Live shopping is a timing bet, not a technology bet

At Shoptalk Fall, Yakuel sat on a panel titled “Overhyped: Blunt Takes on Hot or Not Technologies,” where he called live shopping underhyped. Going live is easy, and TikTok Shop already shows U.S. shoppers will buy that way. What’s still rare is a brand that runs it well and profitably. “We have yet to see the real proliferation of this throughout the entire commerce experience here in the U.S.,” he said. His case rests on timing, not doubt about the format itself. He expects brands will need one to two years to build the operational muscle to do live shopping well, which is exactly why he thinks the investment window is now.

He also pushed back on the idea that American shopping culture is too different for live shopping to work here. In his view, cultural differences surface in content preferences and product choice, rather than in basic buying behavior.

Agentic commerce is real, right up until the point of sale

Yakuel drew a sharper line on AI agents that shop on a customer’s behalf, a theme running through this year’s keynote lineup. He has no issue with AI handling customer service, where shoppers may not notice the switch or may even prefer it. The actual purchase is different. “Consumers still want to engage with the brand, play around with the website, really experience it, the digital version of their product, and then make a buying decision,” he said, rather than complete a transaction without ever touching the brand’s site. For now, that’s where he draws the limit on agentic commerce’s reach into conversion.

AI resets what agencies owe their clients

Inside WITHIN, AI isn’t about doing the same work for less. It’s about raising what’s expected of everyone involved. “The bar of expectation goes up for everyone everywhere,” Yakuel said, from what consumers expect of brands to what he expects from his own staff. Brands that use AI to cut headcount are picking the wrong fight, in his view. Their competitors will use the same tools to do more, not less, and there’s no ceiling on the work AI can support.

Last click still runs the boardroom, even though no one defends it

Last-click attribution is the rule Yakuel says should have died years ago, and it hasn’t. He said WITHIN still regularly onboards clients whose boards and leadership measure success that way, long after the method’s flaws became industry consensus. “You cannot grow on top of last click measurement,” he said, arguing it fails to credit the funnel-building and awareness work that drives growth.

He framed the gap between belief and practice as a hypothetical: ask a room how many think last click is bad, and every hand goes up. Ask how many still have to optimize or report on it, and many of the same hands stay up.

TikTok Shop and Amazon aren’t rivals, and neither are your other channels

On where customer attention sits today, Yakuel rejected the idea of choosing between TikTok Shop and Amazon. “We’re talking about the same humans, and we have to treat them like humans, not like channels,” he said, arguing that brands winning on TikTok Shop only sustain that investment because it’s lifting their Amazon performance too.

That logic extends into his advice for 2027 budget planning. Rather than judging a new channel against the average return of an established one, he wants marketers comparing marginal dollars instead. “The question is really about marginal return, not average return,” he said. A lower average ROAS on a new platform can still produce a better overall return, once the least efficient spend elsewhere gets reallocated to it.

Marketers finalizing 2027 plans face a simple trade. Spreading budget across more channels carries less risk than assuming last year’s best dollar will perform the same way.