The number was the beginning of the question
Taco Bell could see that some younger customers were buying fewer drinks with their food. The obvious commercial response might have been a promotion. A different explanation was waiting in the refrigerator at home.
Greg Creed recounts that insight on ChatGTM. In his telling, colleague Tom Wagner identified that these customers could get a soft drink at home without spending their own money on it. They could not get the same Taco Bell food there. Skipping the drink was a sensible decision from the customer's side of the counter.
Creed describes taking that problem to Pepsi and asking for a proprietary beverage. That became the starting point for Baja Blast. The important part of his account is the movement from an observed decline to a reason for it, then to a product idea shaped by that reason. A discount would have answered a different question.
Keep asking after the first plausible answer
Creed distinguishes what data shows from the insight needed to act on it. A transaction report can reveal that something changed. Understanding the customer's situation often requires conversation, including the exchanges frontline employees already have during ordinary service.
He encourages repeated questioning rather than accepting the first explanation that sounds reasonable. That does not mean every problem requires a formal research project. It means the team should be able to explain the behavior it hopes to change without substituting a familiar marketing tactic for an answer.
Take a product with declining attachment. Before changing the price, ask when guests stopped adding it, what else they buy and what alternatives they already have. A convenience problem, a relevance problem and a price problem may produce similar-looking charts. They will not necessarily respond to the same intervention.
The Baja Blast story is particularly useful because it respects the customer's logic. The customer was solving a personal budget problem. Once the business understood that, it could offer a reason to behave differently.
The people closest to the guest need room to act
The interview with Michael Beck and Chad Horn moves from marketing into leadership, but the connection is direct. A company depends on employees to notice what customers need, relay what they learn and make decisions during service. Those actions are shaped by the experience of working there.
Creed describes an earlier Taco Bell practice that allowed frontline employees to resolve customer problems within a defined spending limit. His emphasis is on the permission it created. The team could help without carrying every small issue through a chain of approvals.
The historical dollar limit is less useful to another business than the operating principle. Establish a clear boundary, teach people how to use it and support reasonable decisions made within it. A service promise becomes more credible when the employee delivering it has some authority to keep it.
That also requires listening after the interaction. Repeated service recoveries can reveal a problem the organization should fix upstream. Empowerment works better when individual judgment and organizational learning remain connected.
Leadership changes the weight of an ordinary comment
Creed tells a small story about a leader casually saying a wall might look good blue. Someone may paint it by the next day. The example captures how easily thinking aloud can become an instruction when it comes from the person with authority.
He describes learning that people watch a leader's behavior and words closely. That influence can create unnecessary work or make a value concrete. It depends on whether the leader recognizes the signal being sent.
A useful discipline is to distinguish an idea from a decision. Say when something is a question, when input is welcome and when the team has agreed to act. Otherwise, employees may spend their judgment interpreting the boss instead of improving the customer experience.
Creed also distinguishes being a strong individual contributor from getting the best out of a team. The person leading does not have to supply every good answer. They do need to create conditions in which another person's insight can be heard and developed.
More processing power still needs a better question
Creed is interested in the useful work AI can support, but he is concerned about accepting its outputs without critical thinking. Faster analysis can expand what a team examines. It cannot remove the team's responsibility to challenge an explanation.
That brings the conversation back to the missing drink. A business could process a much larger transaction file and still fail to consider the customer's refrigerator. The opportunity was not concealed in an obscure calculation. It required someone to ask what the customer was doing and why.
Use the data to locate the question. Let conversations sharpen it. Then give the team enough clarity and authority to test an answer. The useful insight is the one that changes what the business does for the person on the other side of the counter.
