Perdue Tested Concept, Pack and Price Before Air Fryer Wings Shipped. The Platform Crossed Nine Figures.

Before Perdue Farms launched its Air Fryer Ready Wings, it tested the concept, the package and the price with consumers. "When we tested our products," Brice Westring, Perdue's senior vice president of R&D, told the room at the PDMA summit in Atlanta this October, "88% of folks said, definitely will buy." The wings also came out first in Perdue's sensory panel against its own benchmarks and its competitors, including Tyson. "That told us we were ready to go."

They were. Westring said that at week 13 the wings were selling at a 29% higher velocity than competitors, and 65% of sales in the category were Perdue's. "As of last year, we crossed the $100 million platform and we continue to grow in this segment."

Perdue was at the summit because it had just won PDMA's 2026 Outstanding Corporate Innovator award, alongside Huntsman Advanced Materials. But the part of Westring's talk that matters most for software teams wasn't the wings. It was a second product he described right after them, a product Perdue made on purpose to sell exactly once.

One system, two speeds

Perdue runs its innovation through a stage-gate process, the kind where a project has to pass a series of checkpoints before it earns more money and more people. What Westring described is a single system with two very different speeds. "One system that can run at two speeds," he said: a full stage-gate for platform bets, and agile sprints for fast-moving trends. The idea is to "use the gates but use them and flex them at the speed that is required based on the risk that you have for each project."

The trend example was a cultural moment, the "six, seven" meme that took over Gen Alpha. "There's nothing special about the product," Westring said. Perdue took nugget molds it already had and made a six and a seven. "It wasn't about how many bags of six, seven nuggets we sold. It was limited. Time we ran one run, that was it." The goal wasn't volume. It was showing up in a moment their youngest customers cared about – "this is how you get that relevancy back."

So in the same company, in the same year, one product got tested on concept, pack and price before launch, and another got exactly one production run with no expectation of scale. Neither one was the wrong amount of rigor. The amount of rigor matched the size of the bet.

What the big bet had to prove

For the platform bet, Westring laid out why the gates keep working for them, and a few of his reasons apply to software almost word for word. "Big bets are broken into small decisions." "Capital discipline, money released only against evidence." "Consumer validation: concept, pack and price tested along the way and validated before we launch." And clear accountability – an approver at every gate, where "anybody can stand up and say we're not ready."

The last one is the part software teams tend to skip. A gate only works if someone in the room is allowed to say "not yet" and be heard, and in a lot of companies the person who'd say it has quietly learned not to. I wrote recently about how most software teams never built that gate at all. Perdue built theirs and then made it explicitly safe for anybody to use.

Westring also named the two reasons that keep the gates from becoming paperwork: portfolio management, so the company is "working on the biggest bets and the things that are going to be the biggest returns," and operational readiness, which he called the do loop – don't launch until the process to make and ship the thing is ready too.

That's a lot of process for chicken wings, and it's easy to read it as big-company caution. But the bet underneath it was real. When Perdue announced the wings in late 2023, it said the opportunity was that 75% of US households had an air fryer and nobody had made a wing specifically for one. A platform means new formulations, packaging, retail placement and supply commitments, all made before anyone knows whether shoppers will choose it. The validation was cheap compared to what it protected.

Software runs every bet at one speed

Most of the software teams I work with have one speed, and it's usually the trend speed applied to everything. A new market, a new pricing model or a platform that will shape the next two years of the roadmap gets the same treatment as a button: write some tickets, build it, ship it and see what happens. The "see what happens" part is fine for a button. It's expensive for a bet that commits a team for quarters.

The opposite failure happens too, in companies that have been burned before. They run every small idea through the full weight of process, with business cases and review meetings for things that would take a few days to build and a few days to undo. That's how teams end up slow on the small stuff and still unvalidated on the big stuff.

The small stuff should move fast. I've always told teams to get their work to 80% and ship it, because people are happier having things done than having them perfect, and a team that ships and learns eventually beats a team polishing for the 95%. I'd also say yes to a small idea that won't hurt the user rather than spend a meeting arguing about it. But that maxim was always about work where the cost of being wrong is small. It was never meant for a platform.

The big stuff deserves Perdue's treatment, scaled to software. Before a team commits quarters to a new direction, someone should test the equivalent of concept, pack and price: whether the problem is real for the people you think have it, whether they'd recognize your answer as theirs, and whether they'd pay what you need to charge. Sometimes that evidence is cheap and fast. The best product validation I ever got was a training quote a client wouldn't pay, and that refusal validated the product before it existed.

How to tell which speed a bet needs

Westring's line about flexing the gates "at the speed that is required based on the risk" is the whole decision rule, and you can ask it about any idea on your roadmap. What does it cost if we're wrong, and how hard is it to undo?

A feature you can build in a couple of weeks, turn off with a flag and forget about if it doesn't land is a trend bet. Ship it, watch it and learn. A new product line, a new customer segment, a rebuilt core architecture or anything that requires hiring is a platform bet. It earns a gate, real evidence before the money moves, and somebody who's allowed to stand up and say "we're not ready."

The real mistake is never deciding which speed a bet deserves. Perdue's team knew the six-seven nuggets were going to be a one-time run before they made them, and they knew the wings would need to prove themselves before anyone scaled them.

Most software teams make that call by accident, if they make it at all.