Matthew Miller
Jul 25, 2014

Q&A: Dip into Darwinism to move markets

SINGAPORE - The principle of 'survival of the fittest' can help companies quickly home in on product concepts and messages that are destined for market success. So claims Affinnova, which employs evolutionary algorithms to assist a client list that includes a who's-who of top FMCG companies. Campaign Asia-Pacific spoke with Johan Vrancken, the company's senior vice president, Asia Pacific.

Affinnova applies genetic algorithms to marketing tasks such as creating product variants
Affinnova applies genetic algorithms to marketing tasks such as creating product variants

The principles of evolution elucidated by Charles Darwin in 1859 have proven useful not only for understanding the origins of species but also for modern-day science and engineering. Computer simulations based on 'survival of the fittest' have helped to create innovations in a wide range of fields including economics, physics and engineering. (For one fun example, check out this NASA antenna designed using the techniques, which sports a design no human engineer would have tried.)

Boston-based Affinnova holds several patents on evolutionary technologies that it applies to marketing problems. The company's client list includes leading FMCG, financial services, life sciences and consumer-electronics companies. Since setting up its APAC HQ in Singapore in 2013, the company has worked with Procter & Gamble, Coca-Cola, Hershey’s, Colgate-Palmolive, AB-Inbev, Nestle, Kellogg’s, Mondelez, Bayer, L’Oreal, General Mills, J&J, Diageo, GSK, Reckitt Benckiser and Mars, according to Johan Vrancken, the company's senior vice president, Asia Pacific.

In this Q&A, Vrancken expands on the company's process, in which consumer feedback plays the role of the environment—deciding which concepts die and which go on to spawn new, improved offspring.

What services do you offer for brands? When and why do brands typically engage with you?

Affinnova offers optimisation technologies that enable brands to bring stronger propositions to the market faster. Brands use us during product development to identify the strongest resonating product concept with target consumers and to optimise the best messaging, communication or positioning strategy for a new product, or a product they are looking to re-stage.  They also use us to optimise the in-market execution of their products, including package design, pricing, sku line-up and promotional strategies.

Johan Vrancken

What’s your secret sauce? 

Affinnova’s patented evolutionary algorithms make us unique. These algorithms allow brand teams to test hundreds, thousands and even millions of ideas with consumers at once and identify in a matter of weeks the strongest proposition or design to bring to market. So unlike market research, which enables a brand team to measure the potential of their ideas with consumers, our technology actually uses consumers to make ideas better. To use an F1 analogy: We don’t just measure the speed of the car; we actually make the engine run faster before we measure it.
Could you elaborate on the evolutionary algorithms? How do they work?

In genetics there is the concept of 'survival of the fittest': Weak genetic combinations die off and strong combinations survive and replicate. Affinnova’s algorithms apply that same principle to the creative variation that brand teams think up for a product idea, messaging or package design. Marketers can develop a concept and include all the creative variations they want to explore. Then the evolutionary algorithms guide consumers through a choice-based exercise where they see one holistic concept at a time and indicate their preference. The algorithm then identifies the creative elements consumers dislike and what they prefer and keeps evolving new holistic concepts until only the strongest options emerge. The fact that consumers are optimising holistic concepts in real-time ensures that all the elements interact and work in harmony together, so that marketers can realistically execute the optimised concept in market and not be stuck with a mutant idea they really can’t act on.  
Can you give examples of successful outcomes? Especially examples where brands were surprised by the outcomes, as it seems one of the presumed benefits is the ability to arrive at conclusions that the brand wouldn’t necessarily think of on its own?

Sprite fielded a packaging design optimisation in five key markets including China. Given Sprite’s established leadership, a new bottle design represented a huge risk to the brand and enormous cost implications. To minimise risk, Sprite needed consumer feedback to optimise the proposed new designs to outperform their current bottle and competitive products. A key challenge was ensuring the new bottle design better conveyed the idea of “refreshment” to a younger audience while maintaining a strong fit with the brand. Five top bottle designs were identified and measured against the current Sprite bottle and key competitors. This process resulted in a winning design that scored strongly in all areas: consumer preference, the ability to convey refreshment and fit with the brand. From start to finish, Sprite received results across five countries in just weeks versus months. As a result of the Affinnova optimisation, Sprite will begin rolling out its new bottle design globally soon.
The most surprising element for new clients especially is the fact that time and time again the top concepts optimised with consumers outperform the marketer’s best guess, which validates why optimisation is becoming an essential stage in any innovation process.  

How can you know that a concept developed through your process would outperform another concept that never actually gets introduced to the market? You don't seem to have a 'control' in the 'experiment'.

We validate performance both from a client and from an Affinnova perspective. In the measurement phase of our process, we compare consumer preference for clients’ optimised concepts against in-market competitor concepts, and this helps us determine the potential impact in market. We not only put in the top concepts coming out of optimisation but also some pre-optimisation concepts, which are essentially our client’s best guesses before they optimise. Traditionally this would have been the one or two concepts tested using basic forms of research. In 99 per cent of the cases our top concepts outperform the pre-optimisation concepts. And on average, optimisation delivers a 50 per cent improvement over the marketer’s best guess.  
In terms of validating our results as a company, we have done five years of market data tracking. What we observed was that Affinnova Optimised products on average generate 2.6 times the first two years revenue versus non-optimised products. This is one of the main reasons why our business has been growing year after year, making us one of the fastest growing companies in our industry.

In what APAC countries and/or sectors do you see your greatest potential?

We have seen a lot of traction within markets like China, Japan, Korea and Australia. Aside from those, we are doing a lot of regional work with teams based in Singapore, Hong Kong and India. Going forward we see a lot of potential in the smaller emerging markets like Vietnam and Myanmar, which are showing increasing interest in our solutions. 



Related Articles

Just Published

2 hours ago

Women to Watch 2022: Elayne Gan, Dow Jones

Recording incredible growth in the pandemic is no mean feat, and Gan boasts a massive 60% revenue growth since taking over the head of sales role in 2019.

12 hours ago

Heritage brands and the power of deliberate creativity

Brands need the courage to step away from the familiar into that fertile territory for innovation.

12 hours ago

Teads and Impact+ partner to evaluate greenhouse ...

Impact+ is a measurement provider that can estimate the carbon footprint of digital communications and suggest how to set up an environmental KPI.

12 hours ago

Haymarket to launch podcasting title PodPod

The editorial brand will cover craft and business of podcasting.