Typical product development pitfalls and how to avoid them

Written by Robert Cassey, Head of Product Development, 28 October 2019

Earlier this year, Robert Cassey, Head of Product Development, attended a Machine Intelligence Garage course on product development as part of our award of £100,000 AWS credits, and membership of the Digital Catapult incubator. He went to London along with a dozen other top UK start-ups for a day of exercises that highlighted the complexities and pitfalls of product development. In this blog post, he shares what he learnt.

It can be tempting to rush into building a product but be warned, too many projects (and start-ups) fail because the core problem at hand wasn’t properly approached. A lot of the mistakes highlighted by Digital Catapult mirrored my own experience with building’s products. Through hard-learned lessons, we’ve found we can execute an intelligent and consistent process of development by sticking to a few basic principles.

Typical product development pitfalls and how to avoid them


1.  Have a clear and focused objective

In the 1990s search engines were contesting for the top spot. Google focused solely on their search with a clean and focused interface, meanwhile, Yahoo muddied the water with content aggregation, news, and weather. This seemingly harmless decision resulted in a userbase less inclined to use their search. After all, Yahoo’s core product came across as fragmented and unfocused. Not a problem for Google in the 1990s.


Google and Yahoo historic sites

In the 1990s battle of search engines, less was more. Image source: Yahoo’s 1990s homepage and Google’s 1990s homepage.

Our primary aim at is to use our AI technology to source the world’s leading specialists and connect them with customers who need their insight. So, to avoid being the Yahoo in this scenario, every time we think of adding new features or making any changes, we ask ourselves whether they’ll help us achieve this aim better. If they don’t, they don’t make it into the product development pipeline. 


2.  Iterate and iterate often

Initially, started with the premise of a self-sign up expert network where key opinion leaders (KOLs) would identify their area of expertise and get funnelled to our customers. Pain points were quickly identified. We had a ‘chicken or the egg’ problem wherein we needed experts to sign themselves up before our customers would consider using our platform, and no expert was going to sign up without the customers to pay them. Furthermore, it transpired more valuable experts, who were worth speaking to, were less inclined to sign themselves up to our platform. Conversely, less knowledgeable experts were signing up in abundance.

It may never feel like a win to go back to the drawing board, but sometimes it’s necessary. It’s easier to swallow the pill of two months of lost work than two years of bad work. For us, a process of iteration and problem solving commenced. We couldn’t rely on self-proclaimed experts, and we needed a way to deliver a global network of experts to our clients. Through a process of iterative evolution, we found directions and approaches that weren’t obvious before and today our AI-driven solution to knowledge exchange is unrecognisably more powerful than what it once was.

A comparison of the Biotechspert and websites

The evolution from Biotechspert to 


3.  Find the right balance between process and progress

At we have something of an evolving relationship with development processes. Too much or too little can slow your team and jeopardise output. Small, fast-moving teams need to land on something of a Goldilocks Zone. The specifics of which don’t matter. It’s simply best to acknowledge that what’s considered best practice isn’t always right in every situation. We use a set of processes and procedures that fit our needs. As we grow, our needs will change and so too must our processes.


4.  Allow (plenty) time for feedback and testing

If you’re doing things right, then testing should be an implicit process that runs in tandem with your regular development. We’ve found by isolating each new feature our team works on, we’re able to quarantine any potential new issues, identify and rectify them before they’re merged into our wider project.

On the other hand, gathering feedback needs to be an active process. At we’ve started a process of internal and external feedback for new features. First, we ensure everything we build meets our internal standards. We then introduce our new features to an external perspective by encouraging a few initial customers eager to interact with our latest tech to test the water first. We’ve found this measured approach to be the best way for us to constantly push boundaries with our AI platform.


So, there you have it, my take on four common product development pitfalls and how has learnt to avoid them. That said, the product development process will always have its hiccups, that’s the nature of the beast and that’s why we love it, right?


If you're looking for an expert and would like to experience the power of our AI technology, send us an expert request.  If you're an expert who'd like to provide insight, get in touch with us too


Topics: Best practice