How are you going to measure your impact?

How are you going to measure your impact?

This is a question we often get as a start-up offering a tool to optimise workforce management in the maritime industry.

And prospective customers are right to ask. For some years now, businesses in all sectors have been looking to digitally transform. The advent of AI – which has brought the productivity benefits of AI into the mainstream - has only accelerated this demand. However, in the current economic climate, businesses need to be clear on the return of any investment – how will this technology improve operations?

The challenge? Organisations that are in the early stages of digitalisation have often relied on legacy systems that were not designed for the age of AI and data analytics. Establishing a set of consistent data points from which to measure the impact of new technology is therefore difficult.

And without these clear key performance indicators (KPIs), it is difficult to demonstrate the value of our product and outline the return on investment for the customer nor build a solid base of business cases to increase chances of future successes.

In other words, how do tech start-ups measure the impact of their product in organisations with legacy systems that weren’t designed to be measured? At Ensemble Analytics, we’ve had the privilege of taking part in various pre-commercialisation R&D and innovation programmes where we worked with some of the UK’s largest port organisations to trial our workforce management software, which uses machine learning to simplify scheduling and improve the efficiency of operations.

We’ve learnt how to bridge the gap between an organisation that wants to improve operations through digitalisation and a tech start-up that needs to demonstrate how their product will improve operations through measurable KPIs.

1) Close collaboration at every level of an organisation

Establishing measurable KPIs must be a collaborative process between the organisation and the start-up. It requires the start-up to get into the weeds of the business, understand how they operate and the insights that matter most. By building a forward-looking framework for performance tracking rather than relying on historical KPIs, it is possible to demonstrate the value delivered in real time, as the customer progressively adopts and engages with the product.

For example, a port manager will be focused on optimising the allocation of their workforce to avoid unnecessary costs of using third-party contractors. For an operative on the ground, their priority will be receiving a guaranteed schedule that aligns with their working constraints while unions will be focused on ensuring fair and safe working schedules. Only when you’ve truly understood the business and its needs can you begin to establish a framework for a ‘good schedule’ (in our case) and set KPIs based on this.

2) Balancing a data-driven approach with years of industry expertise

KPIs cannot be established in a vacuum based on data alone. Nothing beats getting the product into the hands of the people that will use it day in, day out, getting feedback based on their years of experience and knowledge, and establishing KPIs based on their specific needs. Their approval of a product and demand to use it as part of their daily activity is more valuable than any KPI.

At the same time, key decision-makers are still likely to focus on measurable outcomes unless they have the chance to engage with the product themselves, rather than relying only on feedback from day-to-day users. That is why features like reporting tools tailored to strategic needs are crucial when designing the product.

3) Building the case by highlighting single instances of success

When broader baseline data is lacking, an effective approach is to focus on specific, clearly observable cases where the solution delivers better outcomes. These instances, whether it is a day when scheduling ran significantly faster or a shift where resource allocation was particularly effective, can be used as anchor points. From these anchor points, it is feasible to make well-grounded assumptions that allow for extrapolation to predict broader impact.

Even modest extrapolation from a single well-understood example can show a level of improvement that makes the investment case clear. It also allows users and buyers to see the practical implications of the solution without having to wait for full long-term metrics to be in place.

Selling a new technology requires demonstrating its impact. However, this can be challenging when working with organisations reliant on legacy systems that aren't designed for measurement. Bridging this gap demands collaboration across all levels of the organisation, from user feedback to data-driven insights. It involves building the case for impact by highlighting specific, successful examples and using them to draw broader conclusions. Ultimately, transforming an industry and increasing productivity requires input from the entire ecosystem to develop tools that drive measurable change.