How to set a budget when you cannot yet say what a lead is worth

It is hard to set a good budget without knowing what an enquiry is worth, and you cannot learn that without spending something. So set the first budget small and call it a test. Read your own books first. Invoices, margins and old WhatsApp chats hold more than you expect. The proper budget comes after a quarter of writing down every enquiry.

Start with three numbers from your own records

Put these on one page. None needs to be neat.

Average order value and margin. Take twelve months of invoices. Margin means what is left after material, labour and delivery, not the selling price.

From enquiry to quote to order. Scroll back through phone and WhatsApp for a sample of thirty or forty enquiries. How many reached a quote? How many of those became orders? A rough count is enough to start.

Who brought each order. Referral, walk-in, builder, repeat customer. Write what you remember.

What is one enquiry worth?

Multiply, then divide. The figures here are illustrative, to show the method.

A studio finds an average project brings ₹6 lakh and leaves 18 per cent, which is ₹1.08 lakh of margin. Of thirty enquiries remembered from the last year, ten reached a site visit and three became projects. So one enquiry in ten paid off. An enquiry was worth about ₹10,800 of margin. A site visit turned into a project three times in ten and was worth about ₹32,400.

That is the ceiling for what an enquiry or a visit can cost you and still leave something. It is not a target. It tells you that an enquiry at ₹500 leaves room and an enquiry at ₹9,000 leaves almost none, once your own time is counted.

How big is the test pot?

Choose an amount you could spend and lose without changing how you run the business. If losing it would make you cancel a supplier order, it is too big for a test.

Check it against the margin. The pot should be small enough that a handful of orders proves it, yet large enough to buy that handful if things go well. If one order leaves you ₹5,000 and the pot is ₹4 lakh, you are betting, not testing.

Skip the test approach if you already hold a year of clean source records. Then work backwards from orders, as set out in how to decide your marketing budget.

What the test register needs

One row per enquiry, filled when it arrives.

Two columns are easy to forget: how long you took to reply, and the reason for a no. Without them you cannot tell a weak channel from a slow phone. Tracking lead source in your CRM goes further if you want it.

For example: an interior design studio in Kochi

This is an invented case with illustrative figures. A two-person studio has taken all its work from builders and has never run an ad. The owner wants to try Instagram and Google, runs the sum above, and finds the margin lower than assumed, because carpenters' rates were heavier than the quotes allowed.

The owner sets a pot that can be lost and splits it between the two channels. Before starting, one rule is written: a channel that brings no site visit is stopped, and one that brings a visit gets a second slice. Suppose the register after the pot is spent shows Instagram with forty messages and one site visit, mostly people asking the price of a full flat before any plan existed. Google shows nine enquiries and four visits.

The owner had been about to put the larger share on Instagram, and the mistake would have been to judge by message count. The register shows visits, and by that count Google wins, so the second slice goes there. The register also shows that replies to Instagram messages took two days, which is a fault of the studio's, not the channel's. And the win is modest: the four Google visits produce one project of ₹3.5 lakh, below the ₹6 lakh the sum assumed, so the pot has not paid back yet.

Three rules for each channel, set in advance

1. No qualified enquiry after its slice is spent: stop the channel.
2. Qualified enquiries but no orders: look at your replies and your quotes before giving it more.
3. Orders at a cost you can carry: give it the next slice, and only that slice.

Part of the budget should go to answering enquiries well. A test run on a slow phone line tells you about the phone line.

When do you widen the budget?

Widen only when the register answers two questions in your own numbers: how many enquiries a rupee of spend produces, and how many of those become orders. If you still cannot answer them, the next quarter is another test, not an increase. A channel that works in a quiet quarter may behave differently in a festival one, so note the season beside the totals.

When a test is the wrong tool

It fits a business that sells what people already search for. It fits badly when the order is very large and the sale takes a year, as with some capital machinery. A quarter of ads says little there. Judge by qualified conversations and meetings, and agree with whoever helps you what counts as one before spend starts.

Write the three numbers tonight

Pull out the last twelve months of invoices and write the margin, the enquiry-to-quote count and the sources on one page. The pricing page says how our own quotes are built. If you want a second pair of eyes on the numbers, sales and marketing budgeting is the service for it, and you can tell us what you need.