Mars, Incorporated is a multinational manufacturer and service provider, which has businesses that range from its popular confectionery goods, to pet food and veterinary services. It’s a global business that has over 135,000 employees and brings in approximately $45 billion in revenue each year. The organization is undergoing a ‘digital transformation’, which is focused on process modernization and standardization, as part of an ERP migration project. Central to this project’s success is the use of Celonis’ process mining technology, which is being run out of a center of excellence, in the global services IT division of Mars.
TJ Young, Director of Process Intelligence at Mars, was speaking at Celonis’ user event in Munich this week, where he explained that the vendor’s process mining capabilities started being used at the organization in late 2019, leading to the formation of the process intelligence ‘hub’ in the beginning of 2022.
In 2019, Mars’ journey started in its back-office finance division, namely around accounts receivable, looking at invoice processing efficiency, customer invoice deductions, capitalizing on working capital, and understanding where it could capture and make the most of cash recoveries. Young has built a team, hired talent and formed partnerships with external SI partners to really pursue these initial use cases.
But this was just the beginning, as Young explained:
We started to branch away from finance and accounts receivable as our lighthouse to say: this works and it's actually impacting our bottom line. We used that to get the attention from other business units, such as the front office, and to expand into supply chain and operations.
We started to work with our Petcare division in mid-2022, where we started some work around order-to-cash, and some stuff on purchase-to-pay. But we have really leaned in over the past year and a half on order-to-cash, in particular around order management and fulfillment.
Towards the end of 2022, Mars decided that it would need at least 10 different pilot projects - proof of concepts - all around the globe, intentionally within different business units, different functions, and different regions. This was important, so that Young and his team could prove value in multiple areas of the business and not put all of Celonis’ eggs in one basket. He said:
Our assumption was that not every single deployment would work and first impressions matter for new technologies. So we wanted to place a lot of bets, and then see what works. Scale the ones that work, focus on those learnings during the proof of concept phase, so that when we kind of scale the rest of this capability, we're walking in with a few scars in our cheeks, but much more informed.
A confectionery of learnings
During 2023, Young and his team started to take the learnings found in those deployments and scale “the ones that matter”. Again, there was a lot of focus on order-to-cash, supply chain and operations. Germany in particular proved ambitious in its appetite for process intelligence and was often one of the first to say, according to Young, ‘let’s try something new’. And the aim has been all about learning, iterating and scaling the deployments. Young said:
We did that with not just the pet nutrition business, but across snack and food and multi-sales earlier this year. So we took the German deployments that were successful and scaled it to five other countries in Europe - and now we have many regional deployments that are not just single market specific.
Providing an example of how Celonis has proven useful, Young described Mars’ warehouse operations use case, which sits within the customer care group, and is using process mining and the Celonis platform to pull data from multiple sources so that it can better codify decision making around order management, delivery dates and shipping, into the system. Young said:
It’s helping us decide when and where we need to combine truckloads together to be more efficient with how we load trucks, ship products sooner, and also pull in external weather data as well.
For products like chocolate, it matters in terms of the weather when we're shipping parcels. If it's too hot, we can't ship parcels standalone, we have to load them onto temperature controlled trucks.
All of that decision making, we've relied on human middleware, and we have to look at everything, go back into our ERP system, type everything in. There's no reason to have that manual bottleneck upfront. So we're using Celonis to tie together all the pieces, unify the information from the source systems, and then actually convert insights to action.
We're taking all that decision making, handing it off to an RPA digital worker, which is logged back into the ERP and makes it happen. It sends an email saying: ‘hey, I did my work’. So it's completely seamless.
Digital transformation goals
Looking ahead to 2024, Young and his team are going to continue a lot of conversations with each of the functional segments, trying to understand where it can further penetrate the business and scale pilots that were successful. However, even more importantly, Mars is kicking off a significant digital transformation next year, which is a large-scale, end-to-end process transformation and standardization programme, wrapped around an ERP modernization and migration project. Young said:
Where does process mining come in? It’s around the end-to-end process. Our leaders have been very straight with us, very clear that this is not an IT project. This is a business transformation. It's IT enabled, but it's not IT-led. This is not a simple lift and shift. We're gonna rip out and replace our old ERP with a new one. It's really around process and enabled by IT.
Young said that his team has helped form a strategy the ways in which process mining and intelligence can play a role, which shift as Mars goes through its transformation journey. Firstly, he said:
First and foremost, all the process mapping, process analysis and diagnostic work, we can do up front. Demystifying what we do today, so that we can be more informed to design the future. Understanding at a granular level, the complexities that are going on in our environment.
This means walking into design sessions a bit more informed, right? Supplementing the qualitative information that we've gathered from interviews and surveys that we're doing - we're interviewing thousands of people, and having value stream mapping sessions around what we do today - so that we can kind of aspire what the ‘to be future looks like’, across not just your environment, but other best of breed applications and our systems. And of course those processes.
A tool like Celonis helps Mars back up these qualitative conversations with data and quantify the pain points it needs to address. Young added:
If we can quantify it, we can benchmark it, we can benchmark ourselves against the external market, as well as internally between different regions, different markets, and across business units. If we can benchmark it we can also validate our business case.
There are a lot of big assumptions that are going into this, so anything we can do to ground ourselves in reality now will give leadership the confidence that this business case is real, and can start to paint a more data-driven picture of things like…associate productivity. How many hours are we gonna free up? How much capacity are we going free up with our current workforce, so that we can do more tomorrow than we can today?
Young said that using the real data behind its source systems with Celonis helps Mars inform the business case. In addition to this, it helps the organization for an informed ‘systems gap analysis’, where it can assess where it is today against standard, out of the box modules and processes. Knowing how big that gap is will help it inform its design and deployment plans - but most importantly, according to Young, inform the change management plan. He said:
How tough is this going to be on our associates and our partners?
You're not going to get that level of detail, just from talking to people. You'll hear the top maybe 20/50/70%, if you're lucky, but not the nitty gritty. Not the skeletons in the closet. We're in the business of exposing all of those skeletons in the closet.
The beauty here is not just focusing on diagnostics and data insights, but really coming forward with a unified plan around insights and action. Can we come forward, partnering with our automation teams, our enterprise integration teams, other DevOps groups within IT, all as one team, so that we're not having separate conversations with the business.
We can say to them: we are going to expose all this visibility across these end-to-end processes, but also come with a plan to take action with it. Without that, we've seen some projects really fall flat.
And finally, Mars is going to use Celonis to continuously monitor how its performing. Young finished by saying:
This is something we want to keep turned on, right? This is not something that's a one time effort. This is not an ad-hoc deployment, a one time health check. The beauty of this versus traditional static process mapping and modeling conversations, is that it's living and breathing.
We have live connections to the source systems. Can I start to track process adherence? Can I track where people are reverting back to old ways of working? And there’s a high chance of doing that, if something's tough, if we don't have an understanding of the day in the life of the associate, they’ll go back to what they know.