As pet food companies collect more data than ever before, many are still struggling to turn information into better business decisions. During Petfood Forum 2026, Iván Franco, founder and independent advisor at Triplethree International, explained how artificial intelligence and advanced analytics can help pet food manufacturers move beyond reporting the past and begin predicting future outcomes.
During his session, "AI-powered analytics for pet food: Price elasticity, forecasting and decision-making in Latin America," Franco presented a practical framework that combines data, predictive models and AI to improve pricing, demand forecasting and portfolio management.
"Today, you have more data than ever, but also less clarity than ever," Franco said. "We have dashboards, tracking systems and market research reports from a variety of sources, but all this data is generating more noise than decision-making."
Speed, scale and continuous recalibration
According to Franco, many pet food companies rely on historical reports and intuition when making commercial decisions. Teams often review past performance, discuss results across departments and then repeat previous strategies without knowing how future changes could affect the business.
"What nobody can answer in such meetings right now are questions like, what happens if we raise our price by 5%? Are we going to lose money or not?" he said. "Or which SKUs are we protecting that contribute the least to the business?"
Franco believes AI's greatest strengths are speed, scale and continuous recalibration. Tasks that once required weeks of analysis and costly external consultants can now be completed in hours. Models that were previously limited to a single product or market can now analyze hundreds of SKUs across multiple channels simultaneously.
"Before AI, if we wanted to run an elasticity model, we had to wait four to six weeks," Franco said. "Today, we can set up a model in just a few hours."
LatAm case studies
Franco shared several case studies from Latin America that demonstrate the business impact of predictive analytics.
One example involved a pet food company that had relied on basic customer segmentation for years using only average spend and price data. By applying a clustering model based on actual purchasing behavior, Franco's team identified six distinct customer groups, including a previously overlooked segment of highly loyal premium buyers.
"The company was considering these buyers as basic consumers," Franco said. "It was all the other way around. They were a high-value segment."
Another case focused on price elasticity during a period of high inflation in Mexico. Despite rising production costs, a pet food company had avoided raising prices for 12 months because executives feared losing sales volume. An elasticity model revealed that premium-product buyers were relatively insensitive to moderate price increases.
Armed with the analysis, the company increased prices on its premium portfolio by 8% to 12%, resulting in an estimated four-point improvement in blended margins while experiencing only a slight decline in volume.
"The commercial team didn't want to raise prices because they assumed they would lose volume," Franco said. "However, there was no data supporting that assumption."
Measuring promotion performance
Franco also discussed how AI can help companies evaluate promotional effectiveness. In one project, his team analyzed whether promotions created true incremental demand or simply shifted future purchases forward. The analysis found that only six of 16 SKUs generated lasting demand growth following promotional activity.
As a result, the company reduced its promotional calendar and focused investments on products that demonstrated measurable returns, leading to an approximately six-point recovery in gross margin.
Additional examples will examine portfolio optimization and SKU rationalization, including how predictive models can estimate shopper behavior before products are removed from store shelves.
The bottom line
Franco said the ultimate goal is not to generate more reports or dashboards but to provide executives with clear recommendations that directly affect profitability.
"You don't need more dashboards," he said. "You need decisions."
For pet food manufacturers looking to begin their analytics journey, Franco recommends identifying the business decision that is currently creating the greatest financial risk, leveraging existing data and applying predictive models to generate actionable answers.
"The pet food industry of the future will not be defined by who has more data," Franco said. "It will be defined by who makes better decisions, who makes them faster and who makes them with greater accuracy."
