Weekly Cost Control Reports the Owner Will Read

Every project has a version of this scene: the owner asks "where are we on money?", the engineer opens a 14-tab spreadsheet, and twenty minutes later the owner knows less than when he asked. So he stops asking — until the day the money runs out, at which point everyone discovers the project was heading over budget for four months and nobody said so in a sentence.

A cost report has one job: answer "will we finish inside the budget, and if not, by how much and what are we doing about it?" — on one page, every week, in numbers the reader can check. Everything else is appendix.

The number most sites do not track: committed cost

Sites track spent (bills paid) religiously, and spent is the most misleading number in cost control. The moment you issue a PO for Rs. 14 lakh of steel, that money is gone from your decision space — even though not a rupee has been "spent." A project can show 40 percent spent against 40 percent budget and already be fatally over-committed on the remaining work.

So the report runs on five numbers per cost head:

• Budget — the approved cost plan for this head (Module 5).
• Committed — every issued PO and signed work order, billed or not, plus bills already paid on this head.
• Yet to commit — your current best estimate of everything on this head not yet ordered. This is the column that requires judgment; the others are clerical.
• Forecast final = Committed + Yet to commit.
• Variance = Budget - Forecast final. Negative means over.

The forecast is the headline. Spent tells you about the past; the forecast tells the owner about the ending.

Worked example: G+1 duplex, week 21

A 2,400 sq ft G+1 duplex, budget Rs. 55.2 lakh (about Rs. 2,300 per sq ft — indicative mid-2026 planning figure for a mid-spec build; city and specification move this a lot). Structure is done, masonry underway, finishes partly ordered. All figures in Rs. lakh:

One line for the owner: "Forecast Rs. 56.1 lakh against Rs. 55.2 budget — Rs. 0.9 lakh (1.6 percent) over, driven by steel rates and labour extras; two actions proposed below." That sentence is the report. Note what made it possible: the labour head shows fully committed (12.0 of 12.0) plus extras — a spent-only report would still be showing comfortable headroom there.

The variance block: top three, cause, action, owner

Below the table, exactly three variances explained — not every line, or nobody reads any of them. Each gets one line of cause and one of action:

1. Labour -0.8: extra items signed beyond the lumpsum (bathroom waterproofing redone, compound wall raised). Action: no further extras without written owner approval against a rate quoted in advance; extras register attached.
2. Structure -0.6: steel averaged Rs. 58.4 per kg landed against Rs. 56 budgeted; quantity was on estimate (reconciliation, Lesson 5, ran 4.2 percent — inside norm). Action: none available — rate movement is sunk; head closed, watch released.
3. Finishes +0.3: tile package landed under budget via the comparative statement (Lesson 3). Action: hold the gain as buffer against the sanitaryware package still to commit; do not release it to upgrades yet.

That last point is a discipline of its own: an under-run on an open head is a provision, not a windfall. The fastest way to convert a 1.6 percent overrun into a 5 percent one is to "spend the savings" while later heads are still unbought.

The labour head deserves one more paragraph, because it is where the biggest residential dispute lives. The worked example uses a lumpsum thekedar contract (naka daily-wage gangs would instead show as weekly committed cost from the muster). Either way, the extras register is the dispute shield: every extra item entered with scope, a rate quoted before execution, and the owner's written approval — a saved WhatsApp confirmation is commonly treated as sufficient between the parties, though your contract governs. The classic end-of-project fight — the thekedar's final bill arriving with Rs. 1 to 1.5 lakh of remembered extras, half of them genuinely done, none of them priced in advance — simply cannot happen on a site where extras exist only in the register. RERA forced developers into a documentation regime; the weekly report and its extras register bring the same protection to owner-built projects, where no regulator will do it for you.

The rest of the page

• Cash forecast, two weeks: payments falling due (matched invoices plus retention releases plus wage cycles) against funds available. Cost control and cash planning are different questions; the owner needs both, three lines will do.
• Decisions needed, with dates: "Sanitaryware brand by the 28th or tiling gang demobilises." The report is where slow decisions become visible before they become delays.
• Reconciliation flag: one line per money material — "cement variance 2.4 percent, within norm; steel 4.2 percent, within norm." Green ticks build the credibility that makes the red flag believed when it comes.

The weekly routine — 30 minutes, same day every week

Fixed day, fixed rhythm: update committed from the week's POs and work orders (10 minutes — trivial if the PO file from Lesson 4 is current); revisit yet-to-commit only for heads where scope or market moved (10 minutes); write the headline sentence, the three variances, and the decisions block (10 minutes). Send it even when — especially when — the news is boring. "Forecast unchanged, Rs. 56.1" for six straight weeks is precisely what makes the owner trust week seven's bad news, and forecasts that only appear when they worsen teach owners to dread the report and doubt the engineer.

Common mistakes

• Reporting spent against budget. Hides commitments; the overrun surfaces months late, after the POs that caused it are unbreakable.
• Fourteen tabs, no sentence. If the owner cannot find the answer in ten seconds, the report has failed regardless of its accuracy.
• Explaining every variance. Ten explained variances get less attention than three. Materiality is a kindness to the reader.
• Forecast never moves. A forecast pinned to budget all project and jumping 8 percent at the end was not a forecast; it was decoration. Honest drift, early, is the entire product.
• No action column. A variance without an action is news; the report's job is control. Every red number gets a next step and a name.

Closing the module

You now hold the whole chain: the cycle that routes every purchase through documents (Lesson 1), the rate history that arms negotiation (Lesson 2), the landed-cost comparison that picks suppliers on truth (Lesson 3), the PO and GRN discipline that makes deliveries match promises (Lesson 4), the reconciliation that catches leakage inside a month (Lesson 5), and the one-page report that turns it all into decisions (this lesson). None of these documents is clever alone; chained, they are why two identical houses can finish 8 percent apart in cost. The next module takes this control system into billing — measurement books, RA bills, and getting paid for what the record proves.