Understanding the Debate Over Natural Processed Versus Washed Coffees
The Past, Present, and Future of Product Development
The nature of CPG has fundamentally changed. Many new product development processes are flawed. To be competitive, CPG companies need to adopt new systems.
The nature of CPG has fundamentally changed. In the past, companies succeeded by creating products that were reasonably well-liked by a majority of people.
An okay-for-all approach was successful when there was a limited range of products on a physical supermarket shelf, but it doesn’t work anymore. An exploding ecommerce industry means we make purchases in different ways than we used to. There are now hundreds of products to choose from, and customers expect personalization. But product development processes haven’t changed to fit the new landscape.
“The largest barriers to entry in terms of success [used to be] things like scale distribution and price points. Today this no longer can fly.” — Ryan Ahn, VP of Innovation and Application at Gastrograph AI.
If your company still conducts a handful of Central Location Tests to determine product decisions, you’re using old tools in a new world. You’re in danger of getting left behind. To be competitive, CPG companies need to adopt new systems.
The Challenges of Traditional New Product Development
In traditional new product development (NPD), companies don’t have access to a large volume of high-quality data. Without enough consumer information, it’s impossible to consistently develop innovative, likable products for a specific target market. It’s time for a new approach.
Many CPG companies don’t properly test their products with consumers. When they do, they test sparingly — which means they’re excluding the voice of the consumer from their product development.
Bringing a product to market is a series of decisions. To create products that succeed, you need data insights to inform each decision point. Historically, companies saw Central Location Tests (CLTs) as the most effective way of gathering that information. In a CLT, participants taste products in a controlled environment rather than at their homes. The testing is usually conducted in a single day, so you obtain the results quickly compared with Home Use Tests.
Sourcing participants that match the consumer profile you’re targeting is no easy feat. The logistics involved mean you have to plan CLTs far in advance and be prepared to pay $20,000–$50,000 for a 500-person trial. CLTs are a reserved resource. If you’re the person signing off on an expensive test to refine a product, you want to feel confident that it’s going to succeed.
Ryan Ahn, VP of Innovation and Application at Gastrograph AI, says, “Every year, there [are] thousands of products that get launched [with] no consumer testing whatsoever.” That’s a lot of guesswork going into some very big decisions.
Unreliable, Disposable Data
When companies try to get input from consumers, the specific way CLTs are conducted means that the data isn’t very useful. Consumer preferences vary over time, so the data is also only valid for a short period.
Traditional tests are unreliable in that they can’t deal with the messiness of language. Words are subjective. We might taste the same flavor but describe it differently. That means results from CLTs don’t give you an accurate picture of how different consumers experience a product.
Preference varies by demographic. So, the insights you gather from a CLT in one location can’t be applied globally. For example, if you conducted a CLT in France, the data isn’t useful for a product you’re launching in New York. You’ll have to cross your fingers and hope New Yorkers enjoy the product too or conduct a whole new set of tests.
In traditional testing, you can’t reuse the data you get. To measure the responses from participants, the way you have to pose questions is very narrow. For example, tasters are often asked to compare two products or rate how much they enjoy specific elements of one product. Because the questions are so specific, you can’t apply the responses to other queries. Each time you want to get insights on a new product or keep up with changing preferences, you have to pay (and wait) for a whole new test.
The cost of traditional testing limits innovation. Experimenting with new flavors is simply too high risk — “You don't have formulators bringing experimental prototypes to a 500 person CLT,” says Ryan.
To successfully innovate, experimentation needs to be part of your process: explore different flavors, aromas, and textures. No one wants to be the person to spend a chunk of the R&D budget on a wacky new idea, so CLTs typically take place at the refinement stage, when you’ve already done most of the development. So, there’s no chance for experimentation.
When experimentation does happen, it’s within a narrow sandbox. You can’t imagine how something will taste unless you’ve already tried it. The most creative person in the world is still only able to think of flavor combinations based on their experience. In contrast, an AI system can suggest flavor combinations that a human would never have conceived of.
The Latest Development in NPD: AI Flavor Profiling
Gastrograph AI has developed the first-ever artificial intelligence platform for flavor profiling. The system can interpret data from tasting panels and learn about how different people perceive flavor. With this information, the platform can reliably predict what flavors different consumers will like. AI flavor profiling allows you to reduce your time to market. With predictions from the Gastrograph AI system, you can develop products to target those demographics without going through extensive in-person tastings.
Target Precise Consumer Preferences With an Abundance of Data
When you have an abundance of consumer information, you can make data-led decisions. You’re able to create products that target specific audiences — fast.
Nimble startups are beginning to pose a challenge to established brands. An ability to move fast and respond to specific consumer demands is a competitive advantage. Take Caulipower — a disruptive startup founded in 2016. Caulipower makes healthy, gluten-free products. By targeting a specific consumer need, Caulipower became a success. The small company even appeared on Neilsen’s list of Breakthrough Innovations in 2019, alongside household names like Unilever and PepsiCo.
Gastrograph AI allows you to accurately and quickly predict preferences, which means you can develop products with tastes people will love. Because repeat sales are so important for CPG brands, being able to create likable products consumers will keep buying is valuable. As Ryan puts it, “companies that get this right [are] worth billions.”
Access Reliable, Reusable Data
AI systems process a lot of data very quickly. At Gastrograph AI, the way we collect consumer information means it’s more reliable than data gathered from traditional tasting panels. We train our AI models using this real-world data — these models can be used over and over again to make flavor predictions.
When consumers are tasting products for the Gastrograph system, they have a range of words they can use to describe each flavor, aroma, and texture.
Selecting textures in the Gastrograph AI review app
The Gastrograph system accounts for the messiness of language by creating ‘signatures’ for each flavor.
These signatures are made up of different layers and overlap with each other.
The Gastrograph AI system is trained to do the same with flavor. It can identify when a flavor is present, even when people describe it differently. The system can spot the key elements of a flavor signature — like how my phone can detect what I want to write after only a couple of letters.
New consumer data is regularly added to our AI models — they’re updated on a daily basis. So, our system can learn how consumer preferences change over time and keep making accurate predictions.
Create Innovative Flavors, Aromas, and Textures
The most exciting feature of the Gastrograph AI system is that it can come up with completely new combinations of flavors, textures, and aromas. It makes it possible to develop products a person wouldn’t have been able to conceptualize.
Gastrograph AI has a catalog of over a thousand flavor signatures. Our system can run through different combinations at lightning speed and predict how different consumers will respond to them.
When I hit shuffle on Spotify, it plays music I know and love and gives me some suggestions of songs I might like. Those suggestions are based on my listening habits. They’re often pretty good, and I end up discovering new music I like. Gastrograph AI does the same. The models based on consumer data can predict what different demographics will enjoy. Only instead of suggesting a song, it’s a blueprint for a set of flavors.
Embrace the Opportunity of New Technology
Gastrograph AI removes the guesswork from NPD so you can save money and speed up the traditional development process. It’s also an opportunity to innovate. CPG companies can harness the power of AI to develop truly original products people will love.