IndianOil case study: separate volume growth, asset utilisation and earnings

What this case study establishes

This is an analysis of public company documents, not a commissioned client case or an assessment based on confidential plant data. IndianOil’s FY2025-26 performance table provides comparable company-defined operating and financial measures; its chairman’s discussion provides management’s explanation of the year. Together they support careful description and testable questions. They do not by themselves establish the causal profitability of each refinery, product or procurement decision. [S3] [S8]

The first discipline is to freeze the perimeter. A researcher should retain whether a figure describes the corporation, a group, a refinery business, domestic sales or total product sales including gas and petrochemicals. Those labels are not decorative. Comparing a group figure in one year with a standalone figure in the next can create a spurious growth result even when the arithmetic itself is correct.

Read the operating bridge before the profit headline

The measures reveal that different parts of the system moved at different rates. That is a starting point for questions about capacity, product mix, sourcing and operating availability, not a complete explanation. Pipeline throughput includes an activity boundary that must remain distinct from refinery processing. Likewise, sales can include externally sourced or non-refinery products. A database should prevent a user from adding these rows as if they were independent demand categories.

A volume increase is not an earnings attribution

The company’s table reports EBITDA of INR 73,718 crore in FY2025-26 compared with INR 38,060 crore in FY2024-25. That change is much larger in percentage terms than refining throughput growth, but the comparison does not prove why earnings changed. A proper bridge would examine segment results, realised margins, inventory effects, costs, exceptional items and the rest of the integrated portfolio. It should also respect the company’s precise EBITDA definition. [S3]

One tempting shortcut is to divide total corporate EBITDA by refinery tonnes and call the result a refining margin. That creates a neat-looking ratio with a misleading label because the numerator and denominator do not share a business perimeter. The ratio might be presented as a broad corporate diagnostic if carefully named, but it cannot replace disclosed segment economics. Before benchmarking, write down the business activity represented on both sides of every fraction.

Reported utilisation needs a capacity and period bridge

The chairman’s FY2025-26 discussion reports 107.4% of installed capacity for refineries, while the prior-year investor presentation reports 102% of design capacity for FY2024-25. These are company-reported utilisation measures, not proof of identical asset conditions or the sustainable capacity of every refinery. A comparison should examine commissioning, debottlenecking, turnaround schedules and the exact denominator used. [S8] [S4]

High throughput against an established nameplate is not inherently an analytical contradiction. Nameplate, design, effective and available capacity can differ. Conversely, an asset can add capacity and show a lower utilisation ratio while still increasing absolute throughput. The research task is to reconcile the numerator and denominator over time before attributing an operational improvement or deterioration. That discipline also matters when comparing different companies whose disclosure conventions differ.

Do not confuse a quarterly margin with a full-year margin

IndianOil’s FY2024-25 release separately reports a Q4 GRM of$7.85/bbl, while its investor presentation reports the full-year$4.80/bbl figure. Comparing a competitor’s annual margin with the stronger quarter would introduce a period mismatch. Even after periods match, analysts need to check whether inventory effects, benchmark conventions or other adjustments differ. Public reported margins are valuable evidence, but they become misleading when their labels are shortened beyond recognition. [S9] [S4]

For intelligence buyers, the quality test is reproducibility. A useful case study lets another analyst retrieve the same documents, reconstruct the growth rates and see what the evidence does not establish. The resulting questions may be more valuable than a single ranking: where did economic contribution improve, how durable is it, and what additional operating data would change the conclusion?