Aggregate before you harmonize

Ten mills, ten processes that each worked, and no way to compare them

An 8 to 10 week sprint, September 2025. A global pulp and paper producer, around €10B revenue. Part 2 of 4 in Assembly or Acquisition. Part 1 was the map that has to exist before a standard does. Part 3 is the inventory that has to exist before anything is outsourced. Part 4 is the diagnostic that tells the two problems apart.

Every mill checked incoming material its own way and decided on its own when to raise a claim against a supplier. This is not a discipline problem. It is what happens to a group that grew by acquisition, where each site arrived with a process that already worked.

The record states the position directly: there was no unified view of how quality was verified or when claims should be escalated. Documentation was scattered, checks were inconsistent, and informal resolutions were common.

The client wanted two things across 3 countries, 10 mills and 7 material categories: a more efficient process, and enough of a paper trail to satisfy an auditor.

We found €5m a year in cost impact. The record describes it as uncovered: annual losses from poor raw material quality and fragmented processes, concentrated in 4 of the 10 mills. Nothing in the business had ever shown that number, and why it stayed hidden is the subject of this part.

The only people who could run the comparison were the ten people running the mills

Everyone recognized the problem. Staff at all 10 mills could describe their own site's weaknesses without prompting. What no individual had was a view of all ten at once.

A harmonization finding does not exist until somebody puts the ten processes side by side and compares them clause by clause. The people who could do that were each responsible for one of the ten, and for keeping it running. The record gives the constraint without qualification: mill and procurement teams lacked the bandwidth to redesign fragmented practices while managing daily operations.

A second effect compounds the first, and it concerns stability rather than cost. Settling a dispute informally is efficient for the mill and invisible to the group. A quality manager who resolves a disputed delivery over the phone releases the material, protects the supplier relationship, and creates no record. Repeat that across ten mills and you get a group that looks stable on every operational measure and can show an auditor almost nothing.

Each informal resolution also removes a data point. The thinner the record gets, the more sensible informal resolution looks the next time, and the harder it becomes to build a case for fixing anything. The problem does not just persist. It deepens.

Three channels, three owners, and nobody holding the total

The record identifies three channels through which the arrangement generated cost. They fall into three different categories of cost, which is why the total stayed hidden while each piece of it was visible to somebody.

No function is accountable for all three. Each owner sees something that, in its own context, reads as ordinary operating cost rather than a warning. Kept apart they are three tolerable numbers and no case for change. Added together they exceed €5m a year, concentrated in 4 of the 10 mills. That concentration is a finding in its own right, because it turns a ten-site programme into a four-site pilot.

Four figures, and one of them needs reading carefully

Three of the four are counts of work delivered. The fourth needs care. The record says uncovered and identified. It does not say saved, recovered or banked. The engagement finished before anyone started remediation, and we do not restate the figure as a saving.

A group that could show an auditor very little now has a costed case and a four-site pilot

What that enables. The group now has a business case denominated in a number rather than a concern, and it knows which four sites carry that number. It can test a harmonized process where the cost is measurable instead of rolling it out to ten sites on faith. It also has a design under which claims generate records, which is what turns audit readiness from a claim into a trail somebody can follow.

Settle the vocabulary first, then harmonize by material category rather than by site

Agree common definitions before a common workflow. Ten mills cannot follow one workflow while each attaches a different meaning to checked, approved and acceptable. That is why we tackled certificate of analysis usage, inspection practice and claim criteria first. Those terms are the working vocabulary, and settling them costs very little next to everything that depends on them.

Work through the 7 material categories, not the 10 sites. Seven is a narrower front than ten, and a category you harmonize once holds at every mill that buys it. The work compounds instead of repeating.

Record the handshake. Digital handshakes undo the effect of informal resolution. A claim you record is a claim you can count next quarter, which is what lets the group track the cost rather than rediscover it. The audit trail follows from that record.

Pilot at the 4 mills where the cost sits. Those four let the group test a harmonized process against a measured figure instead of an assumption. The other six become the rollout once the effect is known.

Four boundaries, and a CFO will ask about the first one immediately

How much of the €5m is recoverable. We identified the cost. We did not model how much of it survives a harmonized process, and any figure offered before the pilot finishes would have no basis.

Whether those 4 mills are the worst performers or the best record-keepers. They are the sites where we could trace the loss. A mill with thinner documentation reports less and may well carry more.

How suppliers will react. One group specification and a recorded claim process change the negotiating position on both sides. That sat outside the engagement.

Whether audit readiness improves in practice. The future process improves traceability by design. Whether it improves under an actual audit depends on all ten sites adopting it, which is a separate programme on a different timescale.

Structure and clarity across mills in weeks, not something we could have done alone

That is how the Director of Business Development described it.

The second half of that sentence carries the point, and it says something sharper than "this was hard". The group could not have done it from inside, because the only people positioned to compare ten processes were the ten people each keeping one of them running. Part 1 describes the same trade in a different function.

Next, part 3. A €10B forest products group delivered more than 300 finance services every day, and no list of them existed anywhere. Part 3 is the clearest example of the pattern in this series, and the one where describing the work honestly matters most.

The series in one paragraph. Four engagements, one request, two problems. Fleet maintenance, supplier quality and finance services were assembly problems. The evidence already existed, spread across sites in words that did not match, and the client was buying protected time and a method to reconcile it. The machinery catalogue was acquisition. Three of the four things it needed sat with competitors and customers, and no amount of internal time would have produced them. The two questions in part 4 tell the problems apart before you scope anything.

One case record, and the €5m reproduced with the qualification the record applies

The record. All figures come from the SprintlyWorks case record for Standardizing Supplier Quality Workflows Across Sites, September 2025: an 8 to 10 week sprint for a €10 billion global pulp and paper producer, covering 3 countries, more than 30 interviews, 10 mills and 7 material categories.

The qualification. We reproduce the €5m as the record states it: cost impact uncovered, identified as annual losses from poor raw material quality and process fragmentation across 4 mills. It is neither a saving nor a run rate.

What we leave out. We describe the client rather than naming it, and no detail about suppliers, sites, specifications or claim thresholds appears here.

About this work. SprintlyWorks runs 8 to 10 week sprints for industrial companies that need a written account of how their sites work today, a single harmonized process, or a service catalogue, and cannot release the people who would build it. We work at the sites, reconcile what we find into one structure, and hand over something the client can act on.

Running a standardization programme, or about to scope one? Answer the two questions in part 4 and send us what you get. If it is assembly work, we will tell you what an internal team would need. If it is acquisition work, we will tell you what evidence you are missing.