Start with the arithmetic, not the percentage
Every article on marketing budgets opens with a percentage of revenue. It is the least useful place to start, because it tells you what other companies spend rather than what your business needs.
Start here instead:
- How many new customers do you need this year?
- What does one customer pay you, on average, across their whole relationship?
- What percentage of qualified conversations turn into customers?
- What percentage of leads become qualified conversations?
Work backwards. If you need one hundred customers, close one in four qualified conversations, and qualify one in three leads, you need twelve hundred leads. Multiply by a realistic cost per lead in your category and you have a floor for your acquisition budget. Now the percentage-of-revenue benchmark becomes a sanity check rather than a plan.
The three buckets
Splitting budget by channel names — so much for Meta, so much for Google — hides the important decisions. Split by function instead.
Bucket 1: Demand capture
People already looking for what you sell. Search ads, category search visibility, marketplace presence, maps and local listings, review platforms.
Capture is the highest-return money in most SME categories because you are not persuading anyone to want the thing, only to choose you. It is also finite. Once you own the intent that exists, more spend produces diminishing returns and rising costs.
Weight this heaviest until you have visibly saturated it.
Bucket 2: Conversion and systems
The landing page, the site, tracking, CRM, routing, automation, follow-up sequences, and the proof material sales needs.
This bucket is chronically underfunded because it produces no impressions to screenshot. It also multiplies everything else: a page that converts at four percent instead of two percent doubles the output of every rupee in bucket one, permanently, without raising media costs.
If your conversion rate is unknown or weak, fund this before increasing media.
Bucket 3: Demand creation
Content, social, brand, video, PR, events, and community. Building the intent that bucket one will capture later.
Creation is slower, harder to attribute, and genuinely necessary. Businesses that only ever run capture find their costs rising every year because they are competing for a fixed pool of intent they did nothing to enlarge.
A starting split that works
For a business with real demand and an unproven conversion path:
- Conversion and systems: a heavy front-loaded investment in the first quarter, then a maintenance line.
- Demand capture: the majority of ongoing monthly spend.
- Demand creation: a consistent minority share, protected from being raided in slow months.
- Testing reserve: ten to twenty percent, ring-fenced.
For a business already converting well with saturated capture, shift steadily toward creation. The moment capture costs start climbing without volume gains, you have hit the ceiling of existing intent.
Protect the testing reserve
The most common budgeting failure is spending one hundred percent on what worked last month. It feels responsible and quietly guarantees stagnation, because every channel decays and every creative fatigues.
A ring-fenced testing slice buys you options: a new audience, a new offer framing, a new channel, a new landing page structure. Most tests fail. The ones that work fund the next two years.
Rules that keep testing honest: one variable at a time, a defined success threshold before launch, a fixed budget cap, and a decision date. Tests without decision dates become permanent line items.
Seasonality and cash flow
Indian SME categories often have sharp seasonal shapes — festival cycles, exam calendars, wedding seasons, fiscal year end, monsoon effects on construction and events.
Two rules help:
- Build creative and landing pages before the season, not during it. Peak season is for spending, not for building.
- Do not go completely dark in the off season. Zero spend resets your platform learning and your search presence, and re-entry costs more than continuity would have.
What to cut first when money is tight
In order:
- Broad awareness spend with no measurable downstream effect.
- Channels you cannot attribute at all after three months of trying.
- Production polish on assets that no longer get distribution.
What to protect: the conversion path, response speed, and the search presence you already earned. Those are the assets that make recovery fast when budget returns.
Reviewing the allocation
Set a fixed monthly review with three questions:
- What did each bucket produce in qualified pipeline, not activity?
- Which single fix would most improve output without more spend?
- What is the testing reserve currently learning, and when do we decide?
Half an hour, monthly, with the same three questions produces better allocation than an elaborate annual plan nobody revisits.
The summary
Do not start with a percentage. Start with the customers you need and work backwards. Split by function rather than platform. Fund the conversion layer before scaling media, protect a testing reserve, keep creation funded through slow months, and review monthly against pipeline rather than impressions.
Budgets fail less often from being too small than from being spent on the wrong bucket in the wrong order.
