IKEA's pursuit of the perfect order

Data makes you undeniable

Tim Hills, data lead at the largest IKEA retailer, Ingka Group, explains how persistence, perseverance and process discovery are essential components of building a better customer experience.

What's in a name?

Order-to-cash (O2C) is one of the most fundamental and widely recognised business processes. Starting when a customer places an order and ending once the payment has been received, it is covered by industry benchmarks from bodies such as the American Productivity and Quality Center (APQC).

APQC also offers a helpfully succinct definition of the "perfect order" as "flawlessly taking and fulfilling a customer order ... allocating inventory immediately, delivering product on time, and sending an accurate invoice."

So, simple then? Far from it says Tim Hills, process and data insight development manager at Ingka Group, the largest IKEA Group franchisee, which operates 390 of the home furnishings behemoth's 450 full-size stores across 29 countries. After all, O2C crosses the boundaries of sales, fulfilment, finance, customer services, digital, tech, and more. The way things are named and defined is vitally important. Just getting the departments and people to agree on basic definitions is a half the battle.

"To be honest, I would be surprised if it was only half the battle," Hills told Computing at Celonis' Celosphere event in Munich last week. "It takes a lot of time and good-will building, trust building and confidence building to be able to come up with those definitions."

People are resistant to change and suspicious of any intervention that might force them to alter their ways of working (even if ostensibly for the better) or alter the local balance of power, all the more so if the change is perceived as being driven by tech.

"Those functions and departments, it's almost like they are speaking different languages," said Hills.

Hence the need to spend a long time building relationships locally. Interestingly, he added, coming to agreement on what defines a perfect order between countries - where they really do speak different languages - is easier than it is across business silos.

Ingka Group's long-term plan is to incorporate all its business processes into a SAP ERP implementation, which is ongoing. However, due to legacy and bespoke tech blocking free data flows, some processes have proved more resistant to incorporation than others. So, it was decided to pull those areas, which include sales orders, out of the implementation for the time being, obtain insight as to how they work, and optimise them for integration into SAP at a later date.

What's in a number?

It turns out that the number of possible permutations of operations that make up the "simple" O2C process in a large organisation like Ingka is enormous. "It's ten to the power of 53 times the number of atoms in the universe," quipped Hills.

Whatever the actual number, for the perfect order to become standardised, potential permutations must be whittled down to something much more manageable. Enter process mining, and more specifically process discovery. Ingka is using the Celonis Process Intelligence platform, which ingests event logs from potentially any digital system including CRM, BI, ERP, point-of-sale and collaboration tools to build up a real-time picture of what's actually happening in the business at any given moment, highlighting the root causes of any bottlenecks.

"At the moment, the main capability that we're utilising is process discovery, revealing and unleashing what's currently happening," said Hills. "Where is the inconsistency?"

An optional extra is to apply process modelling to the results, to bring the process closer to the ideal scenario, but in Ingka's case it's more about identifying if a process is working as it should, because then it can be extracted and added to those that currently make up the "perfect order" in an iterative way, Hills explained.

At the same time, he added, it's vital to see O2C as an end-to-end process, otherwise fixing bottleneck in one place might just make things worse downstream. For example, faster warehouse picking could overload the distribution function, or, just as likely, create knock-on effects that are much harder to predict.

"How do we make sure that we're making the right decision and that we're really making that overall improvement that we're expecting to make? You run the risk of not being able to do that when you don't have that visibility."

Process transparency is especially important with "large and complex" orders, an example being kitchens which need to bring together multiple suppliers, timelines and touchpoints, and in countries where the tech is less advanced.

Taking all these factors into account, Ingka has developed a "perfect order performance score" (POPS), and aims to increase this metric by an average of 3.3% globally over the coming year while also ensuring each individual country hits its target.

Flat-packing the solution

In keeping with IKEA's tradition of standardisation, the idea is to drive sales and fulfilment processes in all stores as close to the perfect order as possible by minimising waste and chopping unproductive edge cases, particularly in "large and complex" orders.

Another big number: Ingka Group processes 100 million orders per year. This, Hills points out, means that a one-in-a-million fault happens twice a week. So how to prioritise?

The impact on the customer is key, said Hills, with top priority given to measures like first-time-right resolution. "If you don't get the fix right first time that's guaranteed to really upset the customer".

Using process mining to analyse cases where orders have required repeated remediation also creates a virtuous cycle in which the root cause of the original error can be found and remediated, making it less likely to happen again. This positive feedback effect - continuous improvement in action - favours prioritising what can be fixed now over future interventions. Ironically, said Hills, with all the focus on "perfect", sometimes a quick-and-dirty fix is what's needed.

"An actionable decision made now is almost always better than a perfect decision made later. Insights are lovely but it's action that really counts."

Once everyone is sat around the same (elegant, Scandi-minimalist) table and speaking the same language, action becomes possible and real change can happen remarkably quickly, said Hills. The hope is the approach can then be flat-packed and applied to other areas of the business.

"We started where we had the most momentum and motivation, so order-to-cash, and then built out from there. We've also been using this process to build up the data governance and data quality and those kind of things in these other areas."

Asked for his top tip for others embarking on a similar journey, Hills replied. "Persistence and perseverance. And process discovery. When you connect everything together end-to-end you become pretty undeniable very quickly. And then you're discussing the 'what we should do about it', not 'what does it mean?'."