How Much Should You Spend on Digital Marketing?

There is a standard answer to this question — 7 to 10% of revenue — and it is close to useless for a business deciding what to do next month. Here is a more practical way to think about it.

Start from what a customer is worth

Before any budget makes sense, you need one number: what is a new customer worth to you over a year?

Not the value of one sale. The value of the relationship.

• A salon customer who comes every six weeks at ₹800 is worth roughly ₹7,000 a year.
• A dental patient might be worth ₹15,000 in year one and more after.
• A ₹40,000 interior job with no repeat business is worth ₹40,000, once.

Once you know that number, budgets stop being guesswork. If a customer is worth ₹7,000 and you can acquire one for ₹700, you should spend as much as you can produce capacity for. If acquisition costs ₹6,000, something has to change before you spend more.

A realistic starting split

For a business turning over between ₹10 lakh and ₹1 crore a year, a sensible monthly digital budget is roughly:

Foundation — ₹2,000 to ₹8,000/month Hosting, website maintenance, your Google Business Profile, basic analytics. This is not marketing; it is keeping the shopfront lit. It should never be cut.

Content and organic — ₹5,000 to ₹20,000/month Social posts, SEO work, photos, Google Business posts. Slow, compounding, and the cheapest customers you will ever get.

Paid ads — ₹8,000 to ₹50,000/month Fast, measurable, stops the moment you stop paying. Start at the bottom of that range.

Below about ₹8,000 a month in ad spend, most campaigns cannot gather enough data to optimise, and you end up paying for a learning period you never get past. It is usually better to spend nothing on ads and put the money into content than to spend ₹3,000 and conclude "ads do not work".

The 70/20/10 rule

A workable way to divide whatever you have:

• 70% on what is already working. If Google Maps brings you customers, spend there.
• 20% on the obvious next thing. A channel similar to what works.
• 10% on an experiment. Something you expect to fail four times in five.

The mistake is spending 100% on experiments because they are exciting, or 100% on the safe thing and never finding the next channel.

What to measure — and what to ignore

Measure:

• Cost per enquiry, by channel
• What proportion of enquiries become customers
• Cost per customer against that customer's annual value

Ignore:

• Followers
• Impressions
• Reach
• "Engagement", unless you can connect it to an enquiry

An agency reporting on impressions is reporting on something that cannot be spent.

Give it 90 days, then judge

Almost nothing in digital marketing gives a fair reading in under three months.

• Month 1: setting up, gathering data, most costs and few results. Do not panic here.
• Month 2: the first real signal. Which channel produced enquiries, at what cost.
• Month 3: enough data to shift budget towards what worked.

If someone promises results in three weeks, they are either buying traffic that will not convert or describing something that was going to happen anyway.

Where the money is usually wasted

Ads pointing at a bad landing page. The most expensive mistake there is. Fix the page before raising the budget.

Boosting posts. The Instagram "Boost" button is the least efficient way to spend on Meta. Use the ads manager or do not run ads.

Paying for content nobody sees. Twelve beautifully designed posts a month, no distribution, no plan. Fewer posts and some budget to promote them beats more posts shouting into nothing.

Everything at once. Website, SEO, ads, social and email started in the same month means you cannot tell what worked, so you cannot repeat it.

A starting point if you want a number

For a business doing around ₹25 lakh a year:

• ₹4,000/month foundation
• ₹8,000/month content and local SEO
• ₹10,000/month ads, after the first two are in place

About ₹22,000 a month, roughly 1% of annual revenue. Modest, sustainable, and enough to produce a clear answer within a quarter about where to put more.

The point is not the exact figure. It is that the figure should come from what a customer is worth to you — and that you should be able to check, ninety days later, whether it worked.