Circle Opinion

Lessons Learned: Test, learn, repeat. Discover the backbone of digital services

James Maw

As I’ve covered in my previous two ‘lessons learned’ blogs – How Three Major Organisations Used Challenges to Drive Digital Change and Three Ways Our Clients Pre-Empt Their Challenges to be More Proactive  – challenge-driven innovation is at the heart of all great digital transformation projects.

But at the heart of it all lies one thing: test-and-learn. Test and learn programmes are how we work with our clients to ensure new solutions work. They’re also how we help clients innovate and push their boundaries.

There are lots of ways this can work in practice: here are four examples of how our clients used test and learn to achieve very different outcomes.


Footfall can provide critical information for organisations. Not only does it show how peoples’ movements are changing over time, but it can also identify footfall growth which can have significant safety implications.

For one of our clients in the transport sector, an increase in footfall can indicate the need to alter floor layouts or even influence the design of future buildings.

While shift managers already recorded footfall information each day, they were using paper-based processes which were time consuming and highly inefficient – often preventing the managers from carrying out some of their other daily routines.

Working with our client’s dedicated digital innovation division, we helped to build an app that makes it quick and easy to upload footfall data and supporting documents. Through testing each version as we went and using what was learnt for the next iteration, we refined the app so that it became part of the shift managers workflows – even developing it to allow high-resolution photo evidence to be captured, giving far clearer context to the data that was being recorded.

Through this app, the data is now more easily accessible for senior executives. This means the information can be used to make more accurate decisions around public area design, while helping ensure compliance with health and safety legislation.


At CACI, we understand the importance ongoing consultation plays in our client relationships. Rather than working on a job-by-job basis, we prefer to partner with teams over the long term and foster deep relationships that help create greater value for our clients.

When we helped University College London Hospitals NHS Foundations Trust (UCLH) launch and develop its tuberculosis symptom screening app, that was just the first step in our relationship.

Using regular dialogues and feedback workshops to test and learn what improvements needed to be made, we now work closely to make constant iterations of the app based on direct user feedback. This makes the app more responsive to the health trust’s needs.

As a result, the NHS is gaining access to large amounts of data-driven insight – critical information to justify funding in a budget-focused organisation.


Fully managed services can help some organisations to outsource their biggest development challenges, while also gaining new skillsets in the process. They’re often the most effective way to work with an outsourced vendor.

Following a successful redesign of its website environment, dedicated CACI teams continually enhance and build new features for Mitchells & Butlers’ guest facing platform using a test and learn methodology.

For example, the team has developed new functionalities which are helping each of the company’s brands to streamline the customer experience – especially during busy periods where service can be more challenging.

These include:

  • Pre-order – enabling direct POS integration to help customers pre-order food and drink at busy times, for a more efficient service
  • Alternative venue – offering customers a suitable nearby alternative if there’s no availability at their first-choice location
  • Order at table – giving customers the ability to order and pay on their mobiles for food and drink at any location

This evolving website approach has proved a huge success for Mitchells & Butlers. As well as satisfying customers directly, the company’s agility is also helping it continuously improve its operational performance.


For supermarkets, perishable wastage can be a significant and avoidable drain on revenues – not to mention the negative environmental and social impact it has. One of the best ways to help avoid this is by improving stock control accuracy in a way that combines customers’ shopping pattern data with Machine Learning.

When Waitrose & Partners wanted to reduce waste, we applied Machine Learning to accurately map out when customers would most likely buy a variety of perishable products – helping to make sure the right amount of produce was delivered and on display to keep likely wastage to a minimum.

By working out the most effective times of day to reduce the price of stock approaching its “display until” date, Waitrose & Partners has not only been able to reduce its wastage, it’s also helped to eliminate any unnecessarily large discounts before they’re needed.

And because Waitrose & Partners is using Machine Learning, the process is only likely to get more accurate as its systems learn more about customers’ shopping patterns.


While these clients used test and learn programmes to help them optimise established projects, test and learn can be used at any scale, whatever the project. If you can test different variables based on previous results, you’ll likely see positive outcomes.

But test and learn is only part of the story. If you want the complete picture beyond test and learn programmes, stand by for my last blog in the series where I’ll talk about fully managed digital services that are proven to get results.

Contact us now
James Maw