Ferrara uses Conveo to close the gap between concept score and launch sales

If you have ever had to explain why a lukewarm concept score turned into a strong product launch, you already know Ferrara Candy Company's problem.

At a glance

Carla Mahfuz, who runs consumer insights across Ferrara's expansion markets, saw product launches beat their concept scores so often she reversed the method: real shoppers bought the product, opened it on camera, and described it in their own words. Those words became the concept and the ad behind it, and Ferrara built the next launch on a number it could trust.

Ferrara's launches kept beating their concept scores. The test was grading the words.

Carla's remit covers every market Ferrara is expanding into, plus the brands it has acquired along the way, such as Dori in Brazil and CPK in France. Her job is to find out whether consumers will want a new product before it reaches the shelf. For years that meant a standard concept test: describe the product in words, show consumers the idea, score their intent to buy.

Every insights team has a number it quietly no longer trusts. Carla stopped explaining hers away and rebuilt the test.

One pattern kept repeating.

Ferrara would launch, and the product would outperform the score that was supposed to predict it. Nerds Gummy Clusters was the clearest case. Lukewarm on paper before launch, strong at UK retail after it.

A concept that overscores and then flops gets a post-mortem. One that underscores and then wins gets applause, so nobody looks. It is a good problem to have, until marketing builds a launch plan around a number that keeps coming in too low.

You find out after the money is committed. Production volume, media spend and the retail pitch are all set against the score. By the time the product outsells it, the launch has already been sized and paid for.

The gap was costing the insights team influence. Every launch that beat its score made the score, and the team behind it, easier to discount.

And the test cannot tell you what it missed. A concept is a marketer's guess at the words, rated by consumers who have never tasted the product.

Nobody in that loop can say which words a shopper would use after the first bite. When the product wins anyway, you learn the words were wrong and nothing about which words would have worked. For a candy whose appeal is crunch, chew and tang, that is the one thing you needed. Ferrara's writers were guessing at sensory language the product delivers on its own, and the test was grading the guess.

Consumers taste the product first, then write the concept

So Carla turned the test around. That needs a product people can actually buy, so she ran it in the markets where the launch had already happened.

Most platforms ask you to fit your method to the software. Carla brought the method. Conveo ran it in every market at once.

She had piloted the idea in Germany before she met Conveo. What she could not do there was run it live, on camera, in every market at once.

Consumers went to a store, bought the product with their own money, brought it home and opened it in front of a camera. Conveo's AI research assistant walked each of them through the experience live.

Once they had lived it, the concept landed

Consumers judged the concept against something they had just eaten instead of something they had to imagine. Eight in ten said the concept matched the product. Texture led every other attribute, and the language was consistent enough to build on: a clear crunch, then the chewy center, then the flavor developing through the chew. Consumers also described the first bite as immediate rather than gradual, so the creative now leads with that first bite.

The product had been right all along.

A concept score you can size a launch against, instead of one you have to explain afterwards.

Interviewed alone, people describe what they tasted. In a group, they describe what the room agrees on

A focus group with the product in the room was the obvious alternative. Carla ruled it out on validity.

Social pressure rewrites the reaction before a researcher ever hears it. Thirty-minute one-on-one video conversations, each consumer alone with the product and the camera, keep it intact.

If your current method still puts eight consumers in one room, this is the part worth stealing.

Speed got her attention. Rigor got the study.

Ferrara's insights managers had previously tried several providers. Carla met Conveo on the floor at the CRC conference in Chicago, drawn by one promise. Write the questionnaire once in English, let it auto-translate, field it in every country at once.

That promise got the meeting. But handing a live study to an AI research assistant, the automated interviewer that runs the session, needed more than speed.

Conveo's platform is backed by a research team with 209 combined years of research experience. For Carla that was the deciding factor.

A single brief now replaces six months of market-by-market testing

Pre-recruiting consumers market by market used to put a single concept test at a month or more. A single English brief now auto-translates into the five to ten languages a study needs and fields everywhere simultaneously.

Concept tests close in under a week. Multi-country work that ran about six months across markets now takes just one.

One questionnaire and one moderation approach in every country also means the cross-market read holds together when it reaches global leadership, with no footnotes explaining why one market used a different method.

Segmentation still needs a researcher behind the software

The reverse concept test proved the model. Carla has since pushed it into harder territory: a segmentation study in France. Segmentation is more methodologically demanding than a standard concept test, so Conveo's research team stayed involved. They ran the AI research assistant, helped design the questionnaire, and checked that the segments held up.

Carla asked who stood behind the platform before she asked how fast it was.

Faster only counts with real consumers and real researchers behind it

We asked Carla what she would tell another researcher who wants faster results from real consumers. She starts with the team.

Carla now presents this method from conference stages. That is the bar for a study you trust.

She draws the same line between Conveo and synthetic panels.

And she thinks the reversal reaches past concept testing.

Ferrara's concepts are now written from the consumer's own words, in every market, before the launch plan is built around them.

Want consumer language inside every concept, in every market, before you write it? Book a demo