Interview with John Conway, the CEO of iota-ML [2023]

Hi – so I have worked in marketing analytics since those great days in the 00’s with the explosion of all kinds of new data – an exciting time for marketers. New data was being created all the time at an ever-increasing volume. It was an exciting time for data-driven marketers due to the possibility of using new triggers, consuming more data into campaigns, and personalizing more and more. Our company iota-ML was founded after a lightbulb moment during the first Covid lockdown. 

In recent years I became an independent consultant with brands big and small across different industries, and I hit upon the insight that many of the struggles I had seen marketers have 20 years ago, remain today. This was in spite of the huge rise in the MarTech industry. I concluded that the capabilities data-driven marketers really want are: 

• More relevant campaigns (send customers more of what they prefer) 

• To get more value from their 1st party data (realize value from this asset)

• To use analytical methods (do better targeting) 

• To work in an agile way (respond to business needs, quicker)

• To have tools that fit within their own workflow (i.e. don’t need help from IT) 

So we set out to build a system that released marketers from the restrictive data science process to free them up to THINK and PLAN and DO more work that adds genuine incremental value to their businesses whilst making their day job far more fun, fulfilling and rewarding!

The system is a parallel platform that runs alongside data warehouses and marketing platforms. Once the system is built, marketers jump on the system, and run predictive targeting tasks all on their own. One of the key design elements was simplicity: the winning edge we incorporated is that a marketer needs no more skills than the ability to navigate a web page, to use the machine learning workflows in our platform. It is a no-code, no-drag and drop, click only platform. Simple, simple, simple was the goal. So the system is fast, super simple to use, and lets marketers get on with what they really want to be doing: creating and delivering value to highly targeted audiences! 

In ML, data cleansing, feature engineering, predictive model building and evaluation is normally a process completed by a person, who iteratively works through these steps, making corrections or adjustments at each step. It is a famously complex process and is a major reason companies cannot just “DIY” even with the right staff – it is frequently cited that a Data Scientist spends 80% of their time processing data and only 20% of their time on work that fully utilizes their skills. The engineering challenge was that we needed to build each step with automated “guardrails” and corrective capability, since the machine learning engineers would not be with the Marketer at the time of use. This involved a large amount of coding and simulation of test cases to ensure we had built a robust solution.

We have passionate, studious people, who just love data analytics and ML – that is the key. The team is always studying methodologies and thinking about how they could

 apply to customer problems we’re hearing. 

Interestingly, we are all remote workers and have actually only had one meet-up in the company’s history! The Covid lockdowns didn’t help this, but we are geographically diverse, so we just make sure we are talking regularly on calls and using collaboration tools to help keep things moving. We are a small company so being multi-skilled and resilient in the face of challenges are good qualities to have. 

We address it by never taking it – our clients manage data flows and we work with the systems teams and GDPR managers to create data flows that are free of PII and comply with each client’s GDPR protocols. That all said, our CTO is an ex-UK Government Cyber Security specialist so our system was built with security in mind, and complies to ISO27001 standards.  

We are a scale-up company so retain that nimble execution style of a start-up. An example was we were able to build, release and beta test a product with a chatGPT integration ahead of enterprise players in this space. 

Our flagship client is Volvo UK and they have used the system to increase engagement from email campaigns by 3x and introduced iota-ML more widely in the organization as the system behind their ‘culture of experimentation’ initiative. We contributed towards a 3-minute internal communications video briefed to all UK staff, and system requirements and desired use-cases have continued to increase. 

We are really getting into a flow now on onboarding and recently secured deals through channel partners for the first time, along with launching a client in the USA to give us our first step towards international expansion. The more clients the more we can support them with their ML adoption, it is such an exciting time and many brands have no idea where to start, so we want to be a company that helps brands take ML out of PowerPoint decks and into an applied setting. On the product front, I’m seeing more automation and a coming together of certain workflows into a more holistic output.