Caldergate Distribution Group: Anaplan estate (fictional example)

Generated with CodelessOps Estate Review · Review your own estate

4 Anaplan models, 426 line items, 145M cells as exported. Export date unknown. Latest recorded action run 2026-09-21. Analysis generated 2026-11-18.

The 5 most impactful things to do, picked from 45 candidates across all 42 tests.

  1. 1. Stop repeating text-to-item lookups (FINDITEM) on every cell in Caldergate FP&A (1 formula)

    Caldergate FP&A test 1.4

    Why
    Anaplan runs a formula once for every cell of a module. When a formula only turns a text value into a label or a list item (for example a month code into a Time item), and that answer is the same for every cell that shares the code, working it out on every cell of a large grid is wasted effort. Moving the formula into a small helper module that has only the list the answer depends on means it is worked out once per item and simply looked up from there. In Caldergate FP&A this applies to the line item Journal Cost Centre in the module DAT01 Actuals GL: together 5.0M cells and 6.2% of the model's measured calculation effort.
    The line item
    Line itemModuleEffortCellsFormula
    Journal Cost CentreDAT01 Actuals GL6.2%5.0MFINDITEM(Cost Centres, Source Journal)
    Steps
    1. Open the line item Journal Cost Centre in the module DAT01 Actuals GL (6.2% of measured effort) and note which list its value actually changes with (often Time, or one list).
    2. In a development copy of the model, create a small system module with just that list, put the formula there, and point the places that used the old formula at the new one. Keep the old line item until the check below passes.
    3. Compare the results before and after, and read the Calculation Effort column before and after.
    Done when
    Every value that used the result still matches the original, cell for cell, and the measured effort share has fallen.
  2. 2. Fix the totals of percentages that are added up in Caldergate FP&A (1 line item)

    Caldergate FP&A test 4.16

    Why
    A summed percentage is a number nobody intends. The cells at the lowest level are right; every parent, and every quarter and year where the time summary is Sum, is not. Whoever reads the total on a page or an export reads a wrong figure, and a formula that reads the parent carries it on. In Caldergate FP&A this applies to the line item EBITDA Margin in the module CAL07 P&L by Cost Centre. 1 line item formatted as a percentage divides one amount by another and has the summary method Sum.
    The line item
    Line itemModuleEffortCellsFormula
    EBITDA MarginCAL07 P&L by Cost Centre0.0%19.2KDIVIDE(EBITDA, Revenue)
    Steps
    1. Open the first one on a grid with a parent showing and compare the total with numerator total divided by denominator total; then set the summary to Formula, or to None where no total is needed.
    2. Make the change in a development copy first, keep the original for comparison, and compare the outputs that read the affected line items before applying it in production.
    Done when
    After the change the parent equals the parent of the numerator divided by the parent of the denominator; leaf cells are unchanged.
  3. 3. Consider replacing the long chain of IF tests in the line item Forecast Opex in the module CAL03 Opex with a lookup table

    Caldergate FP&A test 4.6

    Why
    A long chain of 'if this then that' tests is hard to maintain: each new case means editing the formula, and the mapping is easier to check when it sits as rows in a small table read with one lookup. Whether the table also calculates faster depends on the engine and the formula (Anaplan documents that IF stops at the first true branch), so treat the performance side as something to measure, not assume. In Caldergate FP&A, the line item Forecast Opex in the module CAL03 Opex carries 16.0% of the model's measured calculation effort, which is why it is worth a trial.
    The line item
    Line itemModuleEffortCellsFormula
    Forecast OpexCAL03 Opex16.0%5.0MIF ITEM(Accounts) = Accounts.'6100 Rent' THEN 'INP02 Opex Drivers'.Rent ELSE IF ITEM(Accounts) = Accounts.'6110 Rates' THEN 'INP02 Opex Drivers'.Business Rates ELSE IF ITEM(Accounts) = Accounts.'6120 Utilities' THEN 'INP02 Opex Drivers'.Utilities ELSE IF ITEM(Accounts) = Accounts.'6200 Travel' THEN 'INP02 Opex Drivers'.Travel ELSE IF ITEM(Accounts) = Accounts.'6210 Subsistence' THEN 'INP02 Opex Drivers'.Subsistence ELSE IF ITEM(Accounts) = Accounts.'6300 Marketing' THEN 'INP02 Opex Drivers'.Marketing Events ELSE IF ITEM(Accounts) = Accounts.'6310 Digital' THEN 'INP02 Opex Drivers'.Marketing Digital ELSE IF ITEM(Accounts) = Accounts.'6400 Software' THEN 'INP02 Opex Drivers'.Software Licences ELSE IF ITEM(Accounts) = Accounts.'6410 Hardware' THEN 'INP02 Opex Drivers'.Hardware ELSE IF ITEM(Accounts) = Accounts.'6500 Professional Fees' THEN 'INP02 Opex Drivers'.Consultancy ELSE IF ITEM(Accounts) = Accounts.'6510 Audit' THEN 'INP02 Opex Drivers'.Audit Fees ELSE IF ITEM(Accounts) = Accounts.'6600 Recruitment' THEN 'INP02 Opex Drivers'.Recruitment Fees ELSE 0
    Steps
    1. In a development copy, load the 'case to value' table (listed below) into a small mapping module keyed by the list the tests refer to.
    2. Replace the chain in the line item Forecast Opex in the module CAL03 Opex with a single lookup on that table; keep the old formula in a note until the check below passes.
    3. Export the line item before and after and confirm every cell is equal; read Calculation Effort before and after and keep the change only if it is at least as good.
    Done when
    Every cell is equal before and after, the table holds every case the chain held, and the measured effort is no worse.
  4. 4. Check the line items that break their neighbours' pattern in Caldergate FP&A (1 line item)

    Caldergate FP&A test 4.17

    Why
    A row that was missed when a block was copied, pasted and edited looks exactly like this, and so does a deliberate exception. The difference is whether anyone can say why. Spreadsheet audits look for this first because it is where a wrong number hides in plain sight. In Caldergate FP&A this applies to the line item Insurance Phased in the module CAL12 Driver Phasing. 1 line item breaks a pattern its neighbours share: in a run of five or more line items with one formula shape, this one has a different shape, no formula, or one reference its neighbours do not share.
    The line item
    Line itemModuleEffortCellsFormula
    Insurance PhasedCAL12 Driver Phasing0.0%19.2KIF 'SYS01 Time Settings'.Current Period? THEN 0 ELSE 'INP02 Opex Drivers'.Insurance
    Steps
    1. For each, ask the owner whether the exception is meant. Where it is, record why in the line item's notes; where it is not, bring it in line with its neighbours in a development copy and compare the outputs that read it.
    2. Record the answer for each one where the next builder will find it (the notes on the line item or module): keep it, with the reason, or change it.
    Done when
    Every listed line item either matches its neighbours or carries a note saying why it does not.
  5. 5. Find out whether anyone still uses the module CAL05 Opex OLD in Caldergate FP&A

    Caldergate FP&A test 2.1

    Why
    A module that no formula reads and no export uses may be left over from earlier work, or it may be read only by pages and views, which the exports do not show. Until someone checks, it can neither be removed nor trusted. In Caldergate FP&A, no formula outside the module CAL05 Opex OLD reads any of its 14 line items and no export reads it; it occupies 70.2M cells and 50.8% of the model's measured calculation effort. That is what it takes up now, not a saving.
    Steps
    1. Ask the page builder whether any page, dashboard, saved view or line item subset still uses 'CAL05 Opex OLD'.
    2. Ask whether another model imports from a saved view on it.
    3. Record the decision: keep, or retire. If retire: make the change in a development copy first, keep the original for comparison, check the owner's key outputs over a full cycle, then sign off.
    Done when
    Every question above has a recorded answer and the owner has signed a keep-or-retire decision.
Everything else that was found (40)

In rank order. Each one opens in the evidence.

#CandidateModelTestWhy it is not on the planEvidence
6Find out whether anything reads these calculated line items in Caldergate FP&A (3 modules)Caldergate FP&A2.3met the bar; ranked below the top 5F10
7Move per-cell text and lookups into small system modules in Caldergate Data Hub (2 line items)Caldergate Data Hub1.3 1.4 1.5 1.6met the bar; ranked below the top 5F16
8Take out the dimensions these formulas do not use in Workforce Planning (5 line items)Workforce Planning1.7met the bar; ranked below the top 5F34
9Take out the dimensions these formulas do not use in Caldergate FP&A (7 line items)Caldergate FP&A1.7met the bar; ranked below the top 5F2
10Turn off summaries nothing reads in Caldergate FP&A (30 line items)Caldergate FP&A4.12met the bar; ranked below the top 5F43
11Look at the five heaviest calculations in Caldergate Data HubCaldergate Data Hub1.1this is an observation of where calculation time goes, not a change with a measured benefit; the five line items hold 1.3M cells (under 50M), so it is not worth someone's time before the actions aboveF4
12Find out whether anyone still uses the module SYS01 Time in Board ReportingBoard Reporting2.1the module holds 72 cells and 2.8% of its model's measured effort, below the 50M cells or 5% of effort at which a consumer check is worth someone's time before the actions aboveF11 F37
13Remove line items that only copy another in Caldergate FP&A (14 copy line items)Caldergate FP&A4.2the footprint is small (under 1% of its model's measured effort and under 1M cells), so the gain is unlikely to repay the workF12
14Point readers at the source, not at a chain of copies in Caldergate FP&A (4 chains)Caldergate FP&A3.2the footprint is small (under 1% of its model's measured effort and under 1M cells), so the gain is unlikely to repay the workF13
15Move per-cell text and lookups into small system modules in Workforce Planning (3 line items)Workforce Planning1.3 1.4 1.5 1.6the footprint is small (under 1% of its model's measured effort and under 1M cells), so the gain is unlikely to repay the workF29
16Split the formulas that add up and look up in one step in Caldergate FP&A (1 formula)Caldergate FP&A1.2the footprint is small (under 1% of its model's measured effort and under 1M cells), so the gain is unlikely to repay the workF19
17Give hard-coded assumptions a named input in Caldergate FP&A (6 formulas)Caldergate FP&A4.7this test ranks low by nature: it is about upkeep (naming, documentation, housekeeping of formulas and actions), not what the model costs or whether a number is rightF21
18Give hard-coded assumptions a named input in Workforce Planning (3 formulas)Workforce Planning4.7this test ranks low by nature: it is about upkeep (naming, documentation, housekeeping of formulas and actions), not what the model costs or whether a number is rightF30
19Remove or document the empty modules in Caldergate FP&A (1 module)Caldergate FP&A4.14this test ranks low by nature: it is about upkeep (naming, documentation, housekeeping of formulas and actions), not what the model costs or whether a number is rightF24
20Replace selections of a fixed period or version in Caldergate FP&A (3 formulas)Caldergate FP&A4.9this test ranks low by nature: it is about upkeep (naming, documentation, housekeeping of formulas and actions), not what the model costs or whether a number is rightF26
21Replace selections of a fixed period or version in Workforce Planning (1 formula)Workforce Planning4.9this test ranks low by nature: it is about upkeep (naming, documentation, housekeeping of formulas and actions), not what the model costs or whether a number is rightF32
22Remove 1 duplicate copy of one calculation in Caldergate FP&ACaldergate FP&A4.1the footprint is small (under 1% of its model's measured effort and under 1M cells), so the gain is unlikely to repay the workF14
23Remove 1 duplicate copy of one calculation in Workforce PlanningWorkforce Planning4.1the footprint is small (under 1% of its model's measured effort and under 1M cells), so the gain is unlikely to repay the workF15
24Document or split the very long formulas in Caldergate FP&A (1 formula)Caldergate FP&A4.8this test ranks low by nature: it is about upkeep (naming, documentation, housekeeping of formulas and actions), not what the model costs or whether a number is rightF25
25Give subsidiary views used in calculation their own module in Caldergate FP&A (1 line item)Caldergate FP&A4.11this test ranks low by nature: it is about upkeep (naming, documentation, housekeeping of formulas and actions), not what the model costs or whether a number is rightF22
26Give subsidiary views used in calculation their own module in Workforce Planning (1 line item)Workforce Planning4.11this test ranks low by nature: it is about upkeep (naming, documentation, housekeeping of formulas and actions), not what the model costs or whether a number is rightF31
27Turn off summaries nothing reads in Workforce Planning (1 line item)Workforce Planning4.12the footprint is small (under 1% of its model's measured effort and under 1M cells), so the gain is unlikely to repay the workF46
28Turn off summaries nothing reads in Board Reporting (5 line items)Board Reporting4.12the footprint is small (under 1% of its model's measured effort and under 1M cells), so the gain is unlikely to repay the workF48
29Turn off summaries nothing reads in Caldergate Data Hub (1 line item)Caldergate Data Hub4.12the footprint is small (under 1% of its model's measured effort and under 1M cells), so the gain is unlikely to repay the workF41
30Confirm what a zero divisor should show in Board Reporting (3 formulas)Board Reporting4.10this test ranks low by nature: it is about upkeep (naming, documentation, housekeeping of formulas and actions), not what the model costs or whether a number is rightF36
31Confirm what a zero divisor should show in Caldergate Data Hub (1 formula)Caldergate Data Hub4.10this test ranks low by nature: it is about upkeep (naming, documentation, housekeeping of formulas and actions), not what the model costs or whether a number is rightF17
32Confirm what a zero divisor should show in Caldergate FP&A (8 formulas)Caldergate FP&A4.10this test ranks low by nature: it is about upkeep (naming, documentation, housekeeping of formulas and actions), not what the model costs or whether a number is rightF27
33Confirm what a zero divisor should show in Workforce Planning (3 formulas)Workforce Planning4.10this test ranks low by nature: it is about upkeep (naming, documentation, housekeeping of formulas and actions), not what the model costs or whether a number is rightF33
34Look at the five heaviest calculations in Board ReportingBoard Reporting1.1this is an observation of where calculation time goes, not a change with a measured benefit; the five line items hold 720 cells (under 50M), so it is not worth someone's time before the actions aboveF9
35Review the modules with more than 50 line items in Caldergate FP&A (1 object)Caldergate FP&A4.13this test ranks low by nature: it is about upkeep (naming, documentation, housekeeping of formulas and actions), not what the model costs or whether a number is rightF23
36Ask how the imports and exports with no recent run are triggered in Board Reporting (1 action)Board Reporting5.1 5.2 5.3this test ranks low by nature: it is about upkeep (naming, documentation, housekeeping of formulas and actions), not what the model costs or whether a number is rightF49
37Ask how the imports and exports with no recent run are triggered in Caldergate Data Hub (3 actions)Caldergate Data Hub5.1 5.2 5.3this test ranks low by nature: it is about upkeep (naming, documentation, housekeeping of formulas and actions), not what the model costs or whether a number is rightF42
38Ask how the imports and exports with no recent run are triggered in Caldergate FP&A (5 actions)Caldergate FP&A5.1 5.2 5.3this test ranks low by nature: it is about upkeep (naming, documentation, housekeeping of formulas and actions), not what the model costs or whether a number is rightF44
39Ask how the imports and exports with no recent run are triggered in Workforce Planning (3 actions)Workforce Planning5.1 5.2 5.3this test ranks low by nature: it is about upkeep (naming, documentation, housekeeping of formulas and actions), not what the model costs or whether a number is rightF47
40Check the targets loaded by more than one import in Caldergate FP&A (1 target)Caldergate FP&A5.4this test ranks low by nature: it is about upkeep (naming, documentation, housekeeping of formulas and actions), not what the model costs or whether a number is rightF45
41Find out whether anyone still uses the module zz Archive - 2021 Cost in Workforce PlanningWorkforce Planning2.1the module holds 13.4K cells, below the 50M cells or 5% of effort at which a consumer check is worth someone's time before the actions aboveF35 F38 F51
42Find out whether the data loaded here is still needed in Caldergate Data Hub (1 module)Caldergate Data Hub2.4the objects hold 11.1K cells, below the 50M cells or 5% of effort at which an investigation is worth someone's time before the actions aboveF39
43Find out whether anyone still uses the module SYS01 Time Settings in Caldergate Data HubCaldergate Data Hub2.1the module holds 180 cells, below the 50M cells or 5% of effort at which a consumer check is worth someone's time before the actions aboveF18 F40
44Check the objects whose names say leftover in Caldergate FP&A (3 objects)Caldergate FP&A2.5the evidence is only inferred from names, so this is a hypothesis to test rather than something to act onF50
45Check the objects whose names say leftover in Workforce Planning (1 object)Workforce Planning2.5the evidence is only inferred from names, so this is a hypothesis to test rather than something to act onF51

42 tests were run on these exports: 35 found something, 7 came back clear, 0 could not run. See every test and its result.

45 candidates were ranked; 10 met the bar for being worth doing and the first 5 are the plan. The order is a hypothesis, not a verdict. How chosen · Coverage · All findings

Every test that found something offers a candidate action; the plan is the first of them in a stated order, and the rest are listed beneath it with the reason. The order weighs what the exports can show (evidence, measured effort, cells, whether a total may be wrong). It cannot weigh what they do not show: which outputs the business relies on, or what a change would cost you. Treat the plan as where to look first.