How to operationalise "what changed this month" without spending 2 days writing

If monthly reporting always ends with a late-night scramble to write the narrative, you are not alone. The commentary often takes longer than the numbers because it is treated like a blank-page writing task-when it should be treated like a repeatable operating process.

The fastest way to cut "2 days writing" down to 60-120 minutes is to stop asking:

"What should we write?"

...and start building a system that answers:

"What materially changed, why, so what, and what are we doing?"

This matters even more in real-estate portfolios with multiple SPVs, where consolidation, standardised reporting categories, real-estate KPIs (NOI, yields, gearing, % capital returned), and investor/board-ready packs depend on consistent logic and clean roll-ups.

Below is a definition -> common mistakes -> best-practice framework you can implement and then reuse every month.

---

Definition: what "what changed this month" should actually do

A good "what changed" section is not a recap of the P&L. It is a short decision narrative that explains:

1. What moved (the 3-7 changes that matter)
2. Why it moved (drivers, not accounting labels)
3. So what (risk/opportunity/impact)
4. What we are doing (actions + owners + timing)
5. What to watch next (forward-looking signals)

If you cannot answer those five things, stakeholders ask follow-ups-and you end up "writing" for two more days anyway.

---

Common mistakes that make commentary slow (and hard to trust)

Mistake 1: Starting from the trial balance

Writing line-by-line ("repairs up, utilities down...") is slow and unreadable. It also invites questions like, "Which asset drove that?"

Fix: write from exceptions and drivers (portfolio -> asset/SPV -> driver), not from GL lines.

Mistake 2: No thresholds = everything feels "important"

If every variance is treated equally, your commentary becomes a 3-page novella.

Fix: define what counts as "material" so you only write about what truly changed.

Mistake 3: Mixing performance with timing, one-offs, and reclasses

Boards and investors hate "good news" that later turns out to be classification changes.

Fix: label every change as one of:

• Operational performance
• Timing (accruals / seasonality)
• One-off / non-recurring
• Reclassification (policy or mapping change)

Mistake 4: Commentary is not connected to actions

If your narrative does not include "what we are doing," it is not board-ready. It is just reporting.

Fix: every material negative change gets an action (or a decision request).

Mistake 5: Definitions drift across SPVs

If NOI, capex vs repairs, or "operating costs" are inconsistent across SPVs, writing becomes political ("which version of NOI are we using?").

Fix: standardise reporting definitions and mappings so the numbers are comparable before you write about them.

---

Best-practice reporting: the "Commentary Engine" you can run every month

Step 1: Decide your narrative "spine" (7 questions, every month)

Most stakeholders care about the same questions in the same order. Use that to your advantage.

Here is a practical set for real estate portfolios:

1. NOI: up/down vs last month and vs budget - what drove it?
2. Occupancy/leasing: changes, expiries, voids - what is the story?
3. Operating costs: any spikes - what is recurring vs one-off?
4. Capex/refurb: progress vs plan - timeline/budget risk?
5. Cash and distributions: liquidity and payout - any constraints?
6. Debt and rates: covenants, maturities, sensitivity - what do we need to do?
7. Risks and next month priorities: what are we watching, who owns it?

These align with the metrics investors typically want (occupancy, NOI, yields, gearing, % capital returned) and the "investor/board-ready pack" structure many real estate teams use.

Step 2: Put thresholds in writing (so you only write about exceptions)

Define thresholds that determine what gets a comment.

Example thresholds (adjust to your portfolio size):

• NOI variance: comment if > GBP X or > Y% vs last month / budget
• Occupancy: comment if change > 1.0pp portfolio-wide or any asset > 2.0pp
• Operating costs: comment if any category > GBP X or > Y% vs run-rate
• Arrears: comment if > GBP X or if any single tenant > Z% of rent
• Cash runway: comment if any SPV < 3 months runway (or your policy)
• Covenant headroom: comment if DSCR/ICR within 10-15% of trigger
• Capex: comment if forecast-to-complete variance > X% or timeline slips > N weeks

Once these thresholds exist, "what changed" becomes a filtered list, not a writing project.

Step 3: Use bridges, not prose (the fastest way to explain variance)

For key metrics like NOI, use a simple bridge that shows drivers:

NOI bridge (example structure)

• NOI last month
• • income change (occupancy, rent, arrears)
• • cost change (repairs, utilities, management fees)
• +/- one-offs
• +/- reclasses
• = NOI this month

Then your "writing" is just a short interpretation of the bridge.

This is also where solid consolidation and mappings matter: the bridge must be comparable across SPVs to be credible.

Step 4: Maintain a "reason code library" (so you are never starting from scratch)

Create a short list of reason codes that cover 80-90% of variance explanations. Example library:

Performance drivers

• OCC: occupancy/void movement
• RENT: rent changes/indexation
• ARREARS: collections timing / arrears movement
• COST: operating cost movement (recurring)

Non-performance drivers

• ONEOFF: non-recurring cost/income
• TIMING: accruals/seasonality timing
• RECLASS: reclass / mapping / presentation change
• FX: currency movement (if relevant)

Now write a one-line template for each reason code. For example:

• OCC: "NOI decreased due to [pp] occupancy decline at [asset], primarily from [cause]; leasing action: [action], target date [date]."
• ONEOFF: "Costs increased due to a one-off [type] expense at [asset]; not expected to recur; treatment: [above/below NOI]."
• RECLASS: "NOI movement includes a classification change: [GBP X] moved from [old category] to [new category] to align with policy; underlying performance is [up/down/flat]."

You have just removed the "blank page" problem.

Step 5: Assign ownership (commentary is a workflow, not a single writer)

Split ownership so writing is fast and accurate:

• Finance lead: NOI bridge, classification, tie-outs
• Asset/ops lead: occupancy, leasing, arrears, operational drivers
• Treasury/finance: cash, debt, covenants, rate sensitivity
• PM/capex lead: refurb progress, timeline and cost risks

One person (usually finance) compiles and edits to a consistent voice-but they are no longer inventing all content.

Step 6: Run a two-pass process with timeboxes

This is the simplest operational change that saves hours.

Pass 1: Draft (45-60 mins)

• Pull exception list (from thresholds)
• Populate bridge bullets
• Use reason code templates
• Leave placeholders where owners need to confirm facts

Pass 2: Review and sign-off (30-60 mins)

• Confirm the "why" for each exception
• Confirm any one-offs/reclasses wording
• Confirm actions/owners/dates
• Final tone and clarity check

If you do not timebox, commentary expands to fill the available time.

---

A "copy/paste" monthly commentary template

Use this structure directly in your monthly pack.

1) Portfolio summary (5-7 bullets)

• Headline: Portfolio NOI [up/down] [GBP X / Y%] vs last month ([up/down] vs budget).
• Driver 1: [Income/occupancy] impact [GBP X], driven by [asset/SPV] due to [cause].
• Driver 2: [Costs] impact [GBP X], primarily [repairs/utilities/etc.] at [asset]; classified as [recurring/one-off].
• One-offs/reclasses: [GBP X] relates to [one-off/reclass]; underlying run-rate performance is [improving/stable/softening].
• Cash/distributions: Portfolio cash [up/down] to [GBP X]; distributions [paid/held] due to [reason].
• Debt/rates: [Key point] (for example, % floating, hedge expiry, maturity); +100 bps sensitivity [GBP X impact] (if used).
• Focus next month: [Top 2-3 actions] with owners [names/teams].

2) Exceptions table (short list)

List the top 5-10 items that breached thresholds:

• Asset/SPV
• Metric impacted (NOI, occupancy, cash, covenant headroom)
• Variance (GBP / %)
• Reason code
• Action + owner + due date

That is your "operationalisation" in one view.

---

Where automation helps (and where humans stay involved)

Once you have the framework, most of the "2 days" can be automated away:

Automate:

• Consolidation across SPVs/entities
• Standardised roll-ups via mappings (so NOI and cost categories are consistent)
• Threshold-based exception lists
• Bridges (NOI bridge, cost bridge, cash movement)
• First-draft narrative bullets ("what changed", "strongest/weakest", "risks to watch")

Humans must still:

• Confirm drivers that are not visible in the GL (leasing context, disputes, strategic voids)
• Decide classification on "hard buckets" (refurb vs repairs, one-offs, policy changes)
• Approve and sign off (accountability matters)
• Own the actions and decisions (board reporting is a decision tool, not a chatbot output)

This is exactly the human-in-the-loop model behind an "AI CFO" narrative layer: machines draft quickly, people validate and own the message.

---

A practical "implement this month" checklist

If you want a fast rollout, do this in order:

1. Agree the 7-question narrative spine (what you always cover)
2. Define thresholds for what requires a comment
3. Create an NOI bridge and a cash summary format (standard every month)
4. Set up a reason code library (10-15 codes max)
5. Assign owners for operational drivers (leasing/capex/debt)
6. Adopt the two-pass timeboxed process
7. Track actions (owner, due date, status) so the narrative closes the loop

---

Closing: the secret is to write less, but explain more

Teams spend two days writing when they try to explain everything. The best reporting teams do the opposite:

• They define what matters,
• build exceptions and bridges,
• use consistent templates,
• and write only what changed-plus what they are doing about it.

When your portfolio data is consolidated across SPVs with standardised mappings and controls, you can operationalise this into a repeatable monthly system-and even generate high-quality first drafts automatically.