Eighty-four applications, four decisions.
A review is easy to describe and hard to picture. This is one estate carried end to end — the capability model, the applications mapped onto it, a verdict on each, and the order the moves have to happen in.
The company is deliberately unnamed and the numbers are illustrative. They are shaped the way mid-market estates usually are after a couple of acquisitions, but there is no client behind them and none of this is a case study.
Manufacturer and distributor, three divisions
~700, across four sites
Two acquisitions in five years, neither fully integrated
84 applications, $4.1M annual run cost
What the business does
Six capabilities, twenty-four beneath them, written the way the business describes its own work. No system names appear yet. Three are marked as capabilities the business actually competes on — the rest simply have to function.
Customer & Demand
$640k- Marketing & lead capture2
- Quoting & pricingCompete3
- Order management4
- Customer service1
Product & Engineering
$180k- Product data2
- Design & drawings1
- Engineering change control—
- Technical documentation1
Supply & Procurement
$210k- Sourcing1
- Purchasing2
- Supplier management—
- Inbound logistics1
Make & Maintain
$520k- Production planning2
- Shop floor executionCompete3
- Quality & compliance1
- Maintenance2
Fulfil & Service
$390k- Warehousing2
- Inventory3
- Outbound logistics1
- Field serviceCompete1
Finance, People & Risk
$1340k- Ledger & close2
- AP & AR3
- Payroll & time2
- Reporting & analytics5
Where the money actually goes
The same estate, read as annual run cost per capability. $4.1M in total, before anyone has argued about a single system.
Finance, people and risk is the largest line in the estate and the least differentiating. That is not automatically wrong — payroll has to run — but it is the first question a board asks once it can see the bars.
A verdict, and what it commits you to
Ten of the eighty-four, chosen because they show the shape. Every verdict carries a commitment somebody has to accept, which is the difference between a decision and a label.
| Application | Value | Condition | Run cost | Verdict | What it commits you to |
|---|---|---|---|---|---|
Division A order entry (ERP module) Order management | High | Fair | $310k | Invest | Funding the enhancement backlog here, rather than letting the money drift to whichever system is loudest this quarter. |
Division B order entry (acquired) Order management | Medium | Poor | $95k | Migrate | Moving 140 customers and eight integrations onto Division A’s ERP before the transition agreement ends. |
Division C order desk (Access database) Order management | Medium | Poor | Never bought | Migrate | Rebuilding six undocumented reports, against a real deadline: the person who wrote it retires in March. |
Quoting tool (spreadsheet and macro) Quoting & pricing | High | Poor | Never bought | Migrate | Documenting pricing rules that exist only in the file and in one person’s head, before anything can replace it. |
Supplier management (shared mailbox) Supplier management | High | None | Never bought | Invest | Buying something new in an exercise everyone expects to only remove things. |
CRM Marketing & lead capture | Medium | Good | $120k | Tolerate | Living with the data quality everyone complains about, and deliberately spending nothing on it. |
Legacy CRM (acquired) Marketing & lead capture | Low | Poor | $48k | Eliminate | Exporting eleven years of history somewhere readable, and telling two long-tenured users no. |
Reporting environment (analyst-built) Reporting & analytics | High | Poor | $26k | Migrate | Rebuilding forty reports, of which nine turn out to be used. |
Second BI tool Reporting & analytics | Low | Good | $88k | Eliminate | Cancelling at the November renewal — which only works if the nine reports have moved by then. |
Maintenance scheduling Maintenance | Medium | Fair | $64k | Tolerate | Accepting manual work-order entry for another two years, with someone owning that choice. |
Funding the enhancement backlog here, rather than letting the money drift to whichever system is loudest this quarter.
Moving 140 customers and eight integrations onto Division A’s ERP before the transition agreement ends.
Rebuilding six undocumented reports, against a real deadline: the person who wrote it retires in March.
Documenting pricing rules that exist only in the file and in one person’s head, before anything can replace it.
Buying something new in an exercise everyone expects to only remove things.
Living with the data quality everyone complains about, and deliberately spending nothing on it.
Exporting eleven years of history somewhere readable, and telling two long-tenured users no.
Rebuilding forty reports, of which nine turn out to be used.
Cancelling at the November renewal — which only works if the nine reports have moved by then.
Accepting manual work-order entry for another two years, with someone owning that choice.
What the map says that the list did not
Every one of these was available in the raw inventory. None of them was visible until the inventory was attached to a capability model.
Three systems, one capability
Order management is served by an ERP module, an acquired system and an Access database. Nobody had compared them, because each one belonged to a different division and every division was certain theirs was the exception.
The capability you compete on is the least supported
Quoting and pricing is where the business wins — faster, more accurate quotes than anyone else in its segment. It runs on a spreadsheet maintained by one person, and it appeared on no application list.
A third of the spend sits where the business does not compete
Finance, people and risk take $1.34M of the $4.1M. The entire make-and-maintain chain, which is what customers are actually paying for, takes $520k. Neither number is wrong on its own; together they are an argument.
Two capabilities have nothing behind them
Supplier management and engineering change control both happen in shared mailboxes. Absent systems never appear on an application inventory, which is exactly why a capability model comes first.
The eliminations do not come first
$136k of annual cost is removable, and both removals are blocked behind migrations. An estate review that reports savings without reporting the sequence is reporting a number nobody can bank.
The order it has to happen in
This is the part that separates a roadmap from a spreadsheet. Most of these moves are constrained by something other than the budget — a renewal window, a retirement, a close, a transition agreement.
The ledger mapping has to be agreed, and that will not happen during year-end close.
Retiring the analyst-built environment and eliminating the second BI tool.
Waits on the reports, and on the November renewal window.
$88k a year, on a date the contract has already set for you.
Nothing blocks it. A March retirement sets the deadline, not a budget cycle.
The single largest key-person risk in the estate.
The pricing rules have to be written down first, and nobody has ever done it.
The capability the business competes on, off a single file.
Eight integrations, and a transition agreement with fourteen months left on it.
$95k a year and eight integrations retired.
Nothing blocks it — and nothing forces it either, which is why it never happens.
A gap that is now a decision rather than an oversight.
Six moves, not eighty-four. The rest of the estate has a verdict and an owner, and most of those verdicts are tolerate — which is a decision to spend nothing, taken deliberately. That is what a review is for. How the review runs.
Start with a 45-minute briefing.
Bring the renewal you cannot answer, the list nobody trusts, or the two systems that do the same job. We will tell you honestly where to start.