• Research from ESMT Berlin identifies the indicators that people look for in an early-stage idea, in order to gage whether it will be successful or not
• Researchers say that these indicators differ when it comes to a later stage idea
• Findings suggest that AI could identify these characteristics too, saving time and man-power for companies

Innovate or die – sadly, that’s the reality for all businesses nowadays. Innovation has certainly always been a necessity to thrive, but in previous years providing a solid reliable service was, quite often, enough for a business to get by. And when it came to trying something new, the idea-product timeline was a lot longer, meaning organisations could afford to spend a little more time researching, testing and developing a product before they sought to launch it.

Unfortunately, in today’s business context, with rapidly adapting technology, such timeframes are no longer conducive to success. In a constant race to keep pace with – let alone gain an edge over – the competition, companies are having to bring more and more innovations to the table at an even quicker rate. To keep up “business as usual” risks becoming outdated, outpaced and redundant.

Add to that the pressure to identify and invest in the right innovations, and the stakes become even higher.

All innovations are sparked by an initial idea, but not all of those ideas go on to be successful, and it’s incredibly difficult to identify the ones that will go on to flourish. Some of the world’s most revolutionary products could have easily been turned down if they were proposed to other people, or if entrepreneurs didn’t have the passion to pursue them – one of the earliest examples being the lightbulb, which was initially dismissed by a scientist from the Stevens Institute of Technology as ‘a conspicuous failure’, and considered much too far-fetched to be of value to society. Whilst on the flip side, its undoubtable that many ideas have been rejected that perhaps, with the right backing, could have been incredibly successful.

So how can business leaders, investors and other stakeholders know which ideas are set to change the world, and which should be avoided? According to research from ESMT Berlin, there are some characteristics to look out for that give some indication of an innovation’s future prospects.

What are the signs for success?

The research, which was conducted by Professor Linus Dahlander, alongside colleagues from Aarhus University and the University of Amsterdam, looked into how open innovation platforms impact on idea generation, and whether the characteristics people used as success indicators differed throughout the various stages of idea generation.

They found that the characteristics deemed as indicators of a successful idea at the early stage of an innovation’s journey is vastly different from those which are indicated as markers for future success at a later stage.

In an innovation’s initial stages, people tend to look to the status of the idea’s creator; the previous successes they’ve have had personally and the carefully crafted idea presentation they’ve put together, to decide whether or not it is worth further attention and investment. However, these characteristics were found to have little bearing on whether the idea would actually live up to expectations. Instead, they’re better utilised as a way to weed out the bad ideas.

Whereas, at a later-stage in idea development, people are more likely to judge whether an idea will be successful based on how popular it already is amongst the public and how quickly traction is growing.

Can AI lend a helping hand?

One innovation that has revolutionised how our industries work is the advancement in AI and machine learning, which in many instances can replicate human effort and make them both more accurate and efficient. Could this also be true when it comes to identifying promising new ideas?

To this end the researchers sought to analyse whether or not a predictive machine learning model could weed out bad ideas at the initial idea generation stage and identify those ideas that would go on to become successful.

To do so, the researchers analysed data from a company that is synonymous with innovation, reinvention and imagination, LEGO.

The data was gathered from LEGO Ideas – an interactive platform where any user can submit a proposal for a new LEGO set, which then goes through various rounds of crowd selection until a winner is found – a premier example of crowdsourcing.

The researchers interviewed a number of LEGO employees as well as reviewing 160 separate ideas from the LEGO platform, to understand what characteristics were similar across the successful ideas. Then, the researchers used a machine learning model to detect any patterns in successful ideas throughout the selection rounds. The aim was to identify whether or not there are specific characteristics that make an idea successful.

Interestingly, the researchers found that it is much easier to predict which ideas the crowd will select at the early-stage of an idea, in comparison to the later-stage.

Certain creator (or ideator) and idea characteristics matter when predicting which ideas are likely to be weeded out in the early stages of crowd selection, meaning that, in the initial stages, decision makers can benefit from a helping hand.

“Our research furnishes innovation managers with a roadmap,” says Professor Dahlander, who, alongside being Professor of Strategy at ESMT is also holder of the Lufthansa Group Chair in Innovation. “By understanding these critical characteristics of successful ideas, managers can sieve out underwhelming concepts right at the inception.”

So, using algorithmic machine learning models to assist in weeding out the bad ideas at a much quicker rate, can afford overwhelmed, time-pressed managers more time to focus on those ideas with the potential for success.

But for the latter stages, human engagement and intellect must still take centre stage. The study found those factors that provided a guide in the early stages less useful for predicting winners at later stages of crowd selection. “The long-term implications? More informed, strategic decisions in innovation management,” continues Professor Dahlander.

As with many things, AI cannot provide the solutions on its own. Human instinct and intellect must still lead the way.

By, Peter Remon

Interested in this topic? You might also like this…