With Torben Fabrin, CIO, Arla
Our all-in approach to the agile way of working has given us a confidence boost that is spreading to the rest of the organization, says Arla’s CIO Torben Fabrin. And perhaps one day, Arla will develop new dairy products in the same way they run IT projects.
From robotics in the administration to virtual assistants ready to chat over Skype, digitalization is everywhere in Arla. Data from a myriad of sources are crunched in an analytics powerhouse, which among other things has cut the time it takes to predict milk intake from three weeks to one hour. IoT sensors alert the supply chain in real-time, allowing them to optimize deliveries before supermarket milk coolers begin to run low. And from the CEO and down, virtual meetings and digital collaboration through Office 365 is the norm.
"Arla want to go digital across our processes, both on the customer side and internally. In 2018, our focus has been on our internal processes. We have cut costs significantly by implementing robotics in various places, especially administration, where a lot of our labor-intensive tasks have been automated at this point. And only a third of those savings are directly from labor cost. Another third comes from a higher degree of standardization and the last third from increased productivity through automation. So, there is a lot of support to digitalize as much as possible," says CIO Torben Fabrin.
The agile express train
Two years ago, Arla made the strategic decision to switch to agile development, and for the past year, all IT projects have been 100% agile. This means that all internal stakeholders and external contractors and consultants must be ready to commit fully to the intensive and highly focused way of working as well as speak the “agile” language. Or, as the saying in Arla goes, according to Torben Fabrin: “Once the agile release train starts moving, you are either on it or under it.”
And according to the metrics, Arla’s agile train is an express train. Since the switch to agile, development time has been cut in half.
"We track time from approved budget to go-live of the first minimum viable product. And since we switched to agile, that time has dropped from an average of 10 months to 4½. Development time is not a new metric, but it gets more attention now than it did before," says Torben Fabrin.
You might suspect that the increase in speed has a negative influence on quality, but according to Torben Fabrin, that’s not the case:
"I find that we are better at getting it right the first time, since all stakeholders (be it colleagues or external customers) are involved from the outset. And if we miss the mark, we realize it faster and can adjust accordingly. The one metric where we are a bit behind is 'deliver as promised'. The budget is fixed, and if we get overly optimistic, there is no extra money. In these situations, it is a big advantage to have the rest of the business closely involved. That means that the other stakeholders usually understand the consequences of the changes and can prioritize from there, if we need to adjust the scope."
Commitment is the key to success
A dedicated sponsor is a necessary requirement for any IT project in Arla, and equal measures of commitment from both IT and the sponsor is the key to success. As an example, Arla seeks to challenge the mindset of fixed status meetings, where a project can be rushed or delayed to suit a stakeholder’s calendar. Instead, the project teams run demos at 14-day intervals, which the stakeholders can attend.
"We never enter projects blindly, because there is a lot of money at stake. So, if we launch a digital project, we need the subject matter experts to be fully focused and ready to have IT dedicate the necessary resources. Where things go wrong is when you are overworked and only able to spare a day of the week to a project. We prefer to limit our staff to one project at a time, and the maximum is two," Torben Fabrin says, and continues:
"The close relationship between sponsor and IT team also rubs off on the adaptation of the agile method throughout Arla. We have developed a common language and gotten a confidence boost that supports the transformation to other areas of the organization. Three of the people who were instrumental to implementing the agile method in IT, are now assigned to our major Calcium-program. If we can successfully export the model, perhaps Arla will one day develop new dairy products in the same way as we run IT projects."
Max budget: 500.000 DKK
For a digital idea to get selected for testing, it needs a strong business owner who believes in the potential. The business owners must dedicate their sharpest subject matter experts to the project for a maximum of three months. That's the framework for all the experiments: A maximum duration of three months and a maximum budget of 500,000 DKK.
"The limited scope and budget make it easier to get approval to rapidly test whether an idea has potential. That also means it is OK to fail. And when we can demonstrate a 15% increase in baggage handling capacity, like we did with our AI project, management is very eager for us to launch even more pilots."
When considering which opportunities to pursue, it's important to prioritize, but also to spot if a rejected tech matures to a viable state. So CPH has set up a scouting function tasked with keeping a tap on the technologies the airport is not actively working with. Like drone technology, which has the potential to disrupt the cargo business, but does not currently fit into CPH’s strategic focus.