Product Development

Why CPG Companies Need Iterative Development and How They Can Do It

In iterative product development, you create a version of your product, test it as soon as possible, refine it, and repeat. A process like this for food and beverages would be extremely difficult if it weren't for a solution like Gastrograph AI.

In iterative product development, you create a version of your product, test it as soon as possible, refine it, and repeat. The iteration process allows tech companies (like Apple and Google) to create innovative, successful products. They take an ambitious idea and keep reworking it based on user feedback until it’s perfect. 

But trying iterative development for food and beverages would be difficult. After all, it’s much easier to test out an app and change some lines of code than to run food tastings and remake a physical product.

One way companies can bring the advantages of iteration to their products is by using Gastrograph AI to gather preference predictions. We’ll help you change the way you approach product development so you can make smart data-backed decisions and create exciting likable products. 

Why Iterative Product Development Works

Iterative product development creates successful products because, at each stage, you have data to shape your decisions. In contrast, in conventional food and beverage development, your product decisions are usually blind. 

conventional product development

iterative product development

CPG companies tend to finalize their products before testing with consumers

In conventional product development, the testing happens so late in the development process that it’s difficult to make big changes. It’d be like trying to change the recipe of a cake that’s already in the oven. With iterative development, because you haven’t spent much time or resources on developing the first version, it's much easier to alter the product in response to feedback. You can keep iterating until your tests tell you that the product is ready for launch.

The iterative process is less risky because you find out early on if an idea is likely to be successful. You don’t waste time, money, or other resources on products that won’t work.

When you bring in the voice of the consumer early in the development process, you’re able to create products consumers actually like rather than what you assume they like. The final product is also likely to be higher quality because there are more opportunities throughout development to make changes to perfect it.

Why You Can’t Iterate with Conventional Methods

In conventional product development for food and beverages, you typically don’t have access to much data to inform product decisions. Physical products are also more expensive to make, so repeatedly changing the product is impractical.

CPG companies typically test their products with central location tests (CLTs), which are so expensive and time-consuming that they’re a reserved resource. Companies have to be selective about when they test. They usually only conduct CLTs toward the end of the product development process to check that finalized products are acceptable for launch.

But testing late in the process means you have limited data to inform your decisions. By the time you get results from a CLT, it’s difficult to make any big changes to your product.

Plus, issues in the way companies conduct CLTs — for example, they collect perception and preference data separately — mean that the data you get isn’t reliable. Brands don’t find out consumers’ true opinions of their products until after they launch. By then, it’s far too late to make any changes — which is one reason why so many products fail.

How to Bring Iterative Development to Food & Beverages with Gastrograph AI

The Gastrograph AI system can reliably predict how consumers will perceive different flavor combinations. We can run different formulations through our system to find out how much different demographics will like them without having to hold multiple in-person tastings.

We give you access to a type of iterative development. You can keep changing the intensity and balance of flavors and predict how consumers will respond to each version. Once you're confident a formula will achieve high preference scores, you can create your final physical product.

Use AI Predictions to Lead Ideation

Gastrograph AI can predict how your target demographic will respond to a product without them having to actually taste it. Instead of testing towards the end of the development process, start checking which flavors are popular with your target audience as soon as possible. Let data lead your ideation.

Our system is able to translate perception data between demographics. For example, we can bring your product onto the system at a tasting in New York and then predict how different consumers all over the world will perceive the same product. 

We suggest exploring the predicted preference scores for flavor combinations at the ideas stage. Dismiss any combinations that get low preference scores, and tap into flavors that’ll be popular with your target demographic.

Alternatively, instead of checking how consumers will respond to your product ideas, use our system to figure out which flavor combinations different demographics will enjoy. Then come up with products that target those preferences. 

We worked with a Japanese beverage company who were looking to create a new carbonated drink. None of the flavor combinations their development team suggested got high predicted preference scores. To help them out, we set our system to find flavors their target demographic would enjoy, and it came up with a suggestion they’d never have thought of — hello, pine soda.

Optimize Flavors

Product testing with Gastrograph AI is easier and more cost-effective than CLTs. You’re able to test after each product change you make without the headache of organizing in-person tastings.

After you’ve settled on the flavor combinations for your product, we’ll help you optimize the balance of flavors to maximize liking. To demonstrate how our optimization system works, we recently ran a study to find the recipe for a perfect Manhattan cocktail

We analyzed the ‘average’ recipe for the cocktail and identified which elements drove preference and which reduced preference.


average v. optimized product development

Our optimized recipe would get higher preference scores

Then, our system ran through hundreds of different formulations to find the one that would perform best with people who enjoy drinking Manhattans. We found the optimal formula for a Manhattan is: 

  • 67.83% whiskey 
  • 32.16% vermouth
  • 1.38% bitters

You’re welcome.

Gastrograph AI can also help you optimize for different locations. We’ll help you pick which version of your product to launch based on your target location. If you’ve got a product that’s successful in one area, we’ll find how you should adjust the recipe to launch it in a new area. 

We recently worked with a snack company that had a spicy corn chip that was popular in Mexico and the USA. Our system assisted them in finding a new version of their recipe so they could launch a successful version of their product in Europe. 

Launch With Confidence 

When you test with Gastrograph throughout your product development, you have the information you need to launch with confidence.

With conventional product development, launching is a big ‘moment of truth.’ You have to wait to find out whether consumers will like your product. “Gastrograph AI removes the guesswork,” says Gen Tsuchikawa, the chief investment manager for Sony Innovation Fund. Improve your product gradually and avoid nasty surprises after you launch.

Our system also helps you understand which attributes drive liking of products. Ryan Ahn, VP of Innovation and Application at Gastrograph, explains companies can “leverage [that] data to better set consumer expectations." 

Design your branding to cue consumers into the flavors they’re most likely to enjoy, or that differentiate your product from competitors. If you know a specific flavor attribute in your product has the biggest impact on preference, you could call it out in your marketing materials or in the product packaging — like how bubly sparkling water matches their can colors to the flavors they want to highlight. 

Think 10x

As well as following their lead on iterative development, CPG companies can also learn from Google’s ‘think 10x’ approach. Think 10x is based on the belief that true innovation happens when you try to improve something by 10 times rather than by 10%. Gastrograph AI means you can test out your ‘think 10x’ ideas without taking risks so that you can dream big, and we’ll provide the data to support you.




Similar posts

Stay up to date with Gastrograph AI

Be the first to know about company updates and industry-related news, straight from our internal subject matter experts.