M&A Value Realization

Synergy rarely fails at the analysis. The workforce line, the procurement line and the systems line were all defensible when they were written. What fails is a translation nobody was assigned: from workforce synergy to which pay calendar survives, whose job architecture the acquired roles map into, and on what date the carrier contract can actually move.

Those decisions are small, unpleasant and individually easy to defer. Defer enough of them and the acquired cost base drifts far enough that the saving cannot be observed at all — not disproved, unobservable. That is the ending most synergy lines get, and it is quiet enough that nobody reports it.

What actually happened

The group was a consumer packaged goods business, roughly $2–3 billion in revenue, private until it went public on the TSX, growing by acquisition. Payroll ran across approximately 21 countries, Workday integrations across more than 20 offices. Polymath’s founder did the human capital work on three bolt-on acquisitions, and years later the diligence and integration planning on a mid-size one.

On the first three he built the workforce models that held mean absolute percentage error to about 3%. The estimate was never the problem. Within weeks of close the company was carrying acquired populations whose pay calendars, eligibility rules, position definitions and approval thresholds matched nothing in its own estate. The model had a line called workforce synergy. Integration received a list of objects, each with a custodian and a renewal date.

The detail is where the money lives. An acquired employee who is not in the position architecture does not forecast alongside anybody else, so their cost sits outside the planning cycle until somebody maps them in. A benefits saving cannot be taken until the plan year turns, and if that date is eleven months out, a year-one synergy was never available.

Some of the value only became visible from the far side. Once the acquired vendor records sat inside the company’s own masters, the team identified and removed roughly $800K of duplicate vendor cost. Real money, in nobody’s synergy line, and not findable until the categories had met the instances.

By the mid-size acquisition the work was arranged differently, deliberately. Diligence and integration planning were one assignment rather than a handoff between teams with different incentives. Before the model was defended, the function named which conversions each line depended on, who would own each, and the date each became possible: carrier renewals, plan years, payroll filing calendars, notice periods.

Underneath both was the same decision, taken late the first time and early the second: make the receiving architecture explicit rather than negotiate it per transaction. Total Rewards was rearchitected across compensation structures, job architecture, benefits and the technology behind them, so an acquired population had somewhere defined to land.

A synergy becomes real on the date an acquired employee, vendor or contract sits inside your own hierarchy. Not before — and no integration tracker records that date.

On a client transaction Polymath names, owns and dates those conversions before the deal model is defended, as it was done the second time rather than the first.

Why it happens

Before close you do not have access, and that is why the deal model looks the way it does. Clean-team restrictions are real, the data room is curated, and the seller has no reason to show you the awkward parts. Categories are the finest resolution available to an outsider, and writing the model at that level is honest.

But value inside an operating company does not sit in categories. It sits in objects: a carrier contract with a renewal date, a vendor master record, a pay calendar, a ledger mapping, the product hierarchy the combined P&L is reported on. Each has a custodian who maintains it as part of an ordinary job, and each changes only on its own cycle.

So the conversion from category to object needs exactly the access you did not have, which means it happens after close, when attention is consumed by keeping two businesses running and clearing obligations that carry legal deadlines.

And every conversion has a loser. Choosing a pay calendar means somebody’s team changes how it works. Choosing a product hierarchy means one commercial organization’s reporting stops matching its own history. Each can be deferred with a reasonable argument, and the person best placed to decide already has a full job. Collectively, they are the entire synergy.

Here is where it becomes invisible. Integration reporting is organized by workstream, and the workstream is moving. Payroll integration is green. The question of which calendar survives sits underneath it, unassigned, with no status of its own. A tracker reports motion. A decision that has not been made looks exactly like one that is in progress.

Then the baseline erodes, which makes the failure permanent. The acquired business keeps operating: it hires, it loses people, it renews contracts, it changes prices. Two or three quarters on, the cost base you undertook to reduce no longer exists in the form you undertook to reduce it from.

So the question stops being answerable rather than answered badly. Did we get the workforce synergy? has no arithmetic behind it once the population, the structures and the accounting hierarchy have moved. An unanswerable question produces no accountability, and none produces no learning.

The strongest objection

The strongest case against this: the integration management office exists precisely for this, and a good one is formidable.

We concede it properly. An IMO is genuinely effective at deadline-bound work: day-one payroll, legal entity readiness, systems access, regulatory filings. Hard dates, public failure modes, and it gets done well.

But an IMO’s authority runs over sequence and status. Whose benefits eligibility rule survives is not a sequencing question. It is a decision belonging to a functional owner who does not report to the IMO and whose own objectives say nothing about the deal. Escalation reaches a steering committee meeting monthly with an agenda full of day-one risk. That committee will not adjudicate a plan-year alignment, so the item returns as an amber flag and stays there.

The second objection is better: you cannot make instance decisions before close, because you often cannot legally see the instances. Correct, and clean-team constraints are not negotiable. But we do not ask you to decide before close. What we propose is to name, before signing, which conversions each synergy line depends on, who will own each, and when each becomes possible — renewal dates, plan years and cutover windows are frequently knowable without confidential data. And the highest-value act takes days: fix the measurement baseline in your own hierarchy at close.

The method

We run six steps. The first three belong to diligence; the last three decide whether the first three mattered.

1. We convert every synergy line into at least one named object. For each line we name the specific thing that must change: a contract, a calendar, a hierarchy, a plan year, a master record. The trade-off we accept: your synergy list gets shorter, because some lines have no object anyone can name — and a line with no object was never a plan. Skip it and integration starts with a number and no instructions.
2. We give each conversion an operating owner, not a workstream. We take the owner to be whoever maintains that object in the ordinary course — the country payroll lead, the category manager, the controller. The trade-off we accept: those people have full jobs, and we are pricing time the deal model never included. Skip it and the conversion belongs to the IMO, which can chase a decision but cannot make one.
3. We date each conversion to the constraint that governs it. Not to the integration plan. To the carrier renewal, the plan year, the payroll period, the notice period, the statutory filing date. The trade-off we accept: several lines move out of year one and the model looks worse than it did at approval. We treat that as the model becoming honest before the board finds out slowly. Skip it and you book value in the year you wanted it.
4. We freeze the measurement baseline in your own hierarchy on day one. We snapshot the acquired cost base mapped into your cost centres, your job architecture and your vendor masters, at close. The trade-off we accept: that mapping is crude on day one and someone will criticize it later for being crude. We take that criticism, because a crude baseline you can reconcile against beats a precise one that no longer exists. Skip it and by quarter three the saving can only be asserted.
5. We set a default convention and require a reason to deviate. We let your convention survive unless there is a documented statutory or contractual reason. The trade-off we accept: you will discard genuinely better practice from the acquired business, and some of it will be better. We take the loss and revisit it deliberately later. Skip it and every convention is negotiated individually, on relationships, for eighteen months.
6. We report synergy at object level in the same pack as the deal-model line. We produce one page: the line, the objects it depends on, which are converted, which are dated, which are open. The trade-off we accept: your integration reporting becomes markedly less flattering in the first two quarters. Skip it and workstream status substitutes for value reporting until somebody external asks.

We do not ask for better diligence. The model was built at the only resolution available, and rebuilding it before close is neither possible nor useful. What we add sits after the model and before the workstream, and it is architectural rather than managerial. In a bolt-on, most of the synergy is Hire-to-Retire work: the acquired population, the calendars they are paid on, the plans they are enrolled in, the vendors serving them. By Foundational Architecture we mean those instances have a defined place to land before they arrive. An integration office can sequence work; it cannot invent a destination, so we define one.

Evidence and measures

Read this against your last transaction. The third column is usually the blank one:

Five measures worth tracking through the first ninety days after close. None requires a new system:

— Deal-model lines carrying at least one named object, a named owner and a date. Expect the first count to be low.

— Proportion of synergy lines whose baseline is expressed in your cost hierarchy rather than the target’s.

— Convention decisions still open at day ninety — pay calendar, plan year, product hierarchy — and the age of the oldest.

— How many items escalated to the steering committee are convention choices rather than risks. That ratio measures whether decision rights were assigned.

— Elapsed time from an acquired employee, vendor or customer entering the estate to appearing in your standard reporting cadence.

If those five move, the deal model is converting into operations. If the only thing that moved is workstream status, it is not.

How Polymath solves this

We start before signing where the timing allows, and in the first fortnight after close where it does not. In that first pass we convert each synergy line into the operating objects it depends on: the position architecture an acquired population maps into, the pay calendar and filing entity that survive in each country, the plan year, the vendor masters. We give each object an owner and an earliest possible date. We deliver that as Foundational Architecture applied to Hire-to-Retire — the destination defined before the population arrives.

We settle the conversions with the country payroll leads, the Total Rewards owner and the group controller who will live with them, then configure them in your own platforms rather than describe them in a plan. We build the day-one baseline the same way, mapping the acquired cost base into your cost centres and job architecture while it still exists in recognizable form. From the second month the object-level report is yours to produce, and we step out of it.

What compounds is the receiving architecture. Your second transaction inherits the mapping rules, the conventions and the baseline method from the first, which turns integration from a bespoke programme into an operating capability: acquired populations enter the planning cycle in weeks rather than quarters, benefits and vendor consolidation have a route and a date, and your next deal model is written with its conversions already named. That is why the second acquisition costs less to integrate than the first.

What it costs to do nothing

The first cost is duplicate running: two carrier contracts, two vendor records, two calendars, two reporting hierarchies. It books as business as usual on the operating P&L rather than against the transaction, which is why it survives. The second is that the purchase price was paid against a model containing those lines. An unconverted synergy is not forgone upside. It is consideration already transferred for a benefit that did not arrive, and it lands on returns rather than on anyone’s operating budget.

The third is the expensive one. Because the loss is never measured, it is never learned. The next transaction is priced with the same category structure, the same confidence and the same absent conversion, at larger scale.

A synergy that never becomes an object with an owner and a date does not fail. It goes missing — and the model that produced it gets used again on the next deal. We start one deal earlier than that, before signing where the timing allows.