TBM after an acquisition: how to merge two IT cost models

In brief

After a deal, two IT cost models rarely add up. Map both to one TBM taxonomy, conform capitalization, and baseline run cost per tower before synergy targets.

Steve Elliott

Steve Elliott · · 7 min read


On this page
  1. Why can't you add the two IT cost models together?
  2. How do you map two cost models to one TBM taxonomy?
  3. What happens when capitalization policies differ?
  4. How do you find application overlap?
  5. Why baseline run cost per tower before setting synergy targets?
  6. How should TSAs show up in the cost model?
  7. What to do on day 1, in the first 100 days and in year 1
  8. Sources

To merge two IT cost models after an acquisition, translate both into one model built on the TBM Taxonomy 5.x, conform the capitalization policy, and baseline run cost per tower in that combined model before anyone commits to an IT synergy number. The order matters: a saving measured against two baselines that count cost differently can't be proven.

This post sets out that sequence, from the taxonomy to transition service agreements, for IT finance leads and enterprise architects who inherit two charts of accounts, two allocation methods and one target. It is not accounting advice; confirm capitalization questions with your controller and auditor.

Why can't you add the two IT cost models together?

Because the totals may describe different things. One company may capitalize internal development labor and the other expense it. One may allocate network cost by headcount, the other by circuits. One may run TBM Taxonomy 4.1, the other 5.0, and a small target may have no taxonomy at all, just a general ledger.

The TBM Council's own example shows how far apart two models can be. When Transport for NSW merged the TBM practices of two agencies that had adopted TBM separately, neither earlier implementation could serve as the starting point because the two agencies operated too differently. The team rebuilt on the TBM Taxonomy as the common framework, ran workshops to agree unified business rules, and finished the merged practice within six months, according to the 2022 TBM Council award case study.

In our view, the realistic default is to build a third model, not to fold one into the other.

How do you map two cost models to one TBM taxonomy?

Use the method the TBM Council published for moving between taxonomy versions. Its July 2025 guide to moving from 4.x to 5.0 recommends a crosswalk that labels each mapping as one-to-one, one-to-many or many-to-one, embeds it in the data pipeline, and tags every object with a version label and mapping source. A merger needs the same crosswalk, with one extra column: which legacy company the line came from.

  1. Pick the target version. We think it should be 5.x even if the acquirer still runs 4.1. Version 5.0 makes Cloud Services its own cost pool, replaces Internal and External Labor with Staffing, retires two towers and adds Data and Smart Devices towers. Migrating once beats migrating twice.
  2. Map each general ledger to cost pools. The Council's 5.0 guide maps External Labor to Staffing > Staff Augmentation and cloud spend booked as outside services to Cloud Services, and advises against catch-all lines such as "Other Operating" unless necessary.
  3. Agree one set of allocation drivers. Where the two companies use different drivers for the same tower, pick one. The Council's fallout guidance favors measurable drivers, such as storage gigabytes for storage, over flat splits.
  4. Run old and new in parallel. The version guide recommends at least one to two reporting cycles of parallel running before cutover. For a merger, that means each legacy model keeps closing until the combined model reconciles to both.
  5. Measure mapped spend by company. The Council defines coverage at each layer as mapped spend divided by total spend, and suggests keeping unmapped cost-pool spend under 1% of total IT financials. Report it separately for each legacy company, so a clean acquirer doesn't hide a poorly mapped target.

What happens when capitalization policies differ?

The acquired company generally moves to the acquirer's policies. PwC says acquirees "need to conform their accounting policies to the acquirer's policies" unless there is significant justification for differences. Software acquired in a business combination is recorded at fair value under ASC 805; after that, internal-use software follows ASC 350-40, according to PwC's software guide.

For the cost model, the risk is a reported saving that is really a policy change. Take an illustrative target that expenses $5 million a year of development labor that the acquirer's capitalization policy would capitalize. Once the policy is conformed, the target's IT opex can fall by up to $5 million in the first year, with amortization of the capitalized cost arriving in later years, and no change in the work. In our view that movement belongs in its own line of the synergy bridge, never in the savings column.

Timing adds a second wrinkle. ASU 2025-06 removes the project stages from internal-use software accounting and is effective for annual periods beginning after December 15, 2027. An integration that runs through 2027 and 2028 could conform the target's policy twice: once to the acquirer's current policy and again on adoption. We think it is worth asking your controller whether early adoption by the combined company would avoid a second policy change. Confirm all of this with your accounting advisers.

How do you find application overlap?

Build one application inventory, then map both portfolios to the same solution and consumer layers. Overlap shows when two CRM systems sit under the same solution with their cost attached.

  • Map applications through products or services. In Taxonomy 5.0, applications are no longer standalone solutions; the Council's guide routes application spend through product or service hierarchies. Use that structure so overlapping applications land under the same product.
  • Classify each application. McKinsey's 2020 article on digital value in M&A sorts function-specific applications into five categories: strategic, leave as is, merge and retire, deprecated, or retire and write off.
  • Judge applications by the deal's sources of value, not cost alone. That is McKinsey's advice for function-specific applications. Cost still matters: Deloitte's October 2023 report says rationalization typically cuts the IT footprint by 20% to 50%, a wide range; your inventory and the deal thesis decide where you land.

Why baseline run cost per tower before setting synergy targets?

Because the published benchmarks count different things, and only your baseline can turn them into a target. Deloitte's 2023 report, based on more than 30 mergers, says IT accounts for over 50% of all synergies. BCG's March 2024 article says technology and data directly drive about 10% of synergies but support up to 85% of business synergies. A 2011 McKinsey article, drawing on its own postmerger work, said 50% to 60% of synergy initiatives were strongly related to IT. These are not contradictions. BCG's 10% counts synergies technology delivers directly. Deloitte's 50%-plus and BCG's 85% count deal synergies that IT enables. McKinsey's 2011 figure counts initiatives, not dollars.

A run rate per tower, in the combined taxonomy, separates those. It shows what both companies spend today, tower by tower, on the same definitions. In our view, targets set against two legacy totals can rarely be tracked.

Expect realization to be slow. EY-Parthenon's August 2024 article says up to 70% of technology synergies in a buy-and-integrate deal arrive 18 to 36 months after close, and puts technology at 2.5% of deal value as a transaction cost driver. In EY's 2024 CIO Sentiment Survey, cited in the same article, only 32% of CIOs said they were significantly meeting deal objectives.

To track realization, we suggest a bridge from the baseline to the current run rate in five lines:

  • Run-rate savings by tower (contracts consolidated, systems retired, staff changes).
  • One-time integration costs, kept out of the run rate.
  • Accounting policy effects, such as the capitalization example above.
  • TSA fees, shown until each TSA exits.
  • Volume changes, such as headcount growth after close.

How should TSAs show up in the cost model?

A transition service agreement is a contract under which the seller keeps providing services, often IT, to a carved-out business after close while the buyer builds its own. Deloitte's 2024 divestiture guide says strategic buyers typically need about 6 to 12 months of TSA support and private equity buyers about 12 to 18. PwC's 2023 article (updated 2025) calls 12 to 24 months the typical exit and argues for 6 to 12. A separate Deloitte article on TSA negotiations says it is not unusual for sellers to add a modest single-digit markup to the actual cost.

We think TSAs need three things in the cost model:

  1. A separate line per TSA, mapped to the tower it stands in for. A TSA for data center hosting maps to the compute and storage towers it replaces, so the baseline shows the cost the buyer will eventually carry itself.
  2. An exit date and the standalone cost that replaces it. The synergy case should show the fee falling away and the replacement cost arriving.
  3. The seller's view. Deloitte's divestiture guide warns that sellers can be left with stranded costs, such as staff who supported the business and don't transfer. On the seller side, the cost model should show which towers lose their consumer when the TSA ends.

What to do on day 1, in the first 100 days and in year 1

PhaseTaxonomy and mappingCapitalizationApplicationsBaseline and synergiesTSAs
Day 1Choose the target taxonomy version and crosswalk formatList policy differences and who decides eachFreeze new purchases of overlapping systems where practicalKeep both legacy models closingRecord every TSA, its fee and its exit date
First 100 daysMap both ledgers to cost pools; agree drivers; start parallel runsAgree conformed policy and its effective date with the controllerBuild the combined inventory and classify each applicationBaseline run cost per tower in the combined model; set targets against itMap each TSA to its towers; plan standalone replacements
Year 1Cut over to the combined model once both legacy models reconcileApply the conformed policy; plan for ASU 2025-06 adoptionStart retire-and-merge decisions for the first waveReport the synergy bridge monthlyTrack exits against plan; on the seller side, size stranded costs

The table is our suggested sequence, not a TBM Council standard.

An acquisition is a test of whether cost, applications and plans sit in one operating model: when they do, each synergy can be traced from the decision to the tower where it shows up.

Sources

Frequently asked questions

Which TBM taxonomy version should a merged company use?

In our view, the current version, TBM Taxonomy 5.x, even if the acquirer still runs 4.1. A merger already forces both companies to remap their ledgers, so moving to 5.x then avoids a second migration later. The TBM Council's version guide recommends a crosswalk that labels each mapping and a parallel run of at least one to two reporting cycles.

Do acquired companies have to adopt the acquirer's capitalization policy?

Generally yes. PwC says acquirees need to conform their accounting policies to the acquirer's unless there is significant justification for differences. Software acquired in a business combination is recorded at fair value under ASC 805, then follows ASC 350-40 if used internally. Confirm how this applies to your deal with your controller and auditor.

How long do IT transition service agreements last?

Published ranges vary. Deloitte says strategic buyers typically need about 6 to 12 months of TSA support and private equity buyers about 12 to 18 months, depending on how entangled the businesses are. PwC describes 12 to 24 months as typical and argues that exiting in 6 to 12 months captures more value.

When are IT synergies realized after an acquisition?

Often 18 to 36 months after close. EY-Parthenon says up to 70% of technology synergies in a buy-and-integrate deal are realized 18 to 36 months after close. In EY's 2024 CIO Sentiment Survey, cited in the same article, only 32% of CIOs said they were significantly meeting deal objectives such as technology synergies, which is why a tower-level baseline matters.

Free playbooks

In-depth guides for product and operations leaders.

Browse the playbooks
Steve Elliott

Steve Elliott

Published November 22, 2026

More on TBM & IT Financial Management

Cost pools, towers, showback, unit cost.