In brief: ROI is simple arithmetic, but clean inputs are the hard part. Count every cost, pick one attribution model and keep it, allow for a 90-day lookback, and judge on cost per customer and margin instead of likes.
Ask ten business owners what their marketing is earning them and most will say “it’s working, I think”. They are not being careless. Digital marketing ROI is hard to pin down, because the money goes out on a Tuesday and the sale lands six weeks later through a phone call nobody logged.
This guide sets out how to work it out properly. What to count, what to ignore, and why your numbers rarely agree with each other.
Key Takeaways
- ROI is simple arithmetic. Getting clean numbers to put into it is the hard part.
- ROAS tells you whether one ad is working, while ROI tells you whether the whole business gained.
- Two tools disagree because they count credit differently, and both can be right.
- Google Analytics can credit a sale to an ad someone clicked up to 90 days earlier.
- Likes and impressions are worth tracking last, if at all.
What Does Digital Marketing ROI Actually Mean?
ROI means return on investment. The sum is straightforward: take what you earned from marketing, subtract what you spent, then divide by what you spent. Spend a lakh, earn three lakh back, and your ROI is 200%.
If the whole idea of measuring spend against outcome is new to you, our guide to performance marketing covers the groundwork this article builds on.
The catch sits in both numbers. “What you earned” needs you to know which sales came from marketing. “What you spent” should include your agency fee, your ad budget, the design work and the hours your own team put in. Most people count the ad budget alone, which makes the answer look far better than it is.
There is a second number people mix up with this one.
ROAS is return on ad spend: revenue divided by ad spend and nothing else, which tells you whether one campaign is pulling its weight.
ROI subtracts every cost you carry, and answers the bigger question of whether the business actually came out ahead.
A campaign can have a healthy ROAS and still lose you money once salaries and software are counted. Both numbers are useful, and they answer different questions.
Which Numbers Actually Matter?
Four, for most businesses, and you can track all of them without expensive software.
Cost per lead (CPL). Total spend divided by the number of leads, which is the quickest honest way to hold two channels side by side.
Cost to acquire a customer (CAC). Total spend divided by the people who actually bought. It lands higher than CPL in any real business, and matters far more, because leads do not pay invoices.
Lifetime value (LTV). What an average customer is worth to you over the whole relationship. A gym member paying every month is worth many times a single sale.
Payback period. How long before a customer has repaid what you spent to win them. If you are spending money you need back this quarter, this is the number that should worry you.
These four are also the numbers any decent lead generation setup should report back without being asked.
Put CAC and LTV side by side and you have the only comparison that really counts. If a customer costs more to win than they will ever spend, no amount of clever creative fixes it.
The market is worth understanding here too. India’s ad spend reached ₹1,55,105 crore in 2025, and digital took 60% of it, or ₹93,156 crore (Madison Advertising Report 2026). Most Indian marketing budgets now sit in channels that can be measured, which makes not measuring them a choice.
Why Do Two Tools Show Different Numbers?
Because they hand out credit differently. This is called attribution, and it explains almost every argument about marketing reports.
Say a customer sees your Instagram reel, searches your name a week later, clicks a Google ad, and buys. Which channel gets the sale? Instagram found them and Google closed them, and you can be sure both tools will claim the credit.
Google Analytics 4 now offers three ways to decide: data-driven attribution, paid and organic last click, and Google paid channels last click. The older options most people learned, first click, linear, time decay and position-based, were withdrawn in November 2023.
One detail catches people out. Google states that “all attribution models exclude direct visits from receiving attribution credit, unless the path to key event consists entirely of direct visits”. So somebody who types your web address straight in has that visit credited to whatever they clicked earlier.
Paid platforms add to the confusion, because each one reports only what it believes it influenced. Our breakdown of what Meta ads actually cost in India goes into how those platform numbers are put together.
The argument we are asked to referee most often is exactly this one: a Meta report and a Google report, both claiming the same sale, with the client wondering who is lying. Usually neither.
None of this means the tools are broken. It means you should pick one model, write down which one, and compare like with like. Changing the model mid-quarter and celebrating the jump is how people fool themselves.
How Long Should You Wait Before Judging a Campaign?
Longer than most people do, and the tool itself tells you why. In Google Analytics 4, acquisition key events, meaning a first visit or a first app open, use a 30-day lookback window by default. Every other key event uses 90 days, adjustable to 30 or 60.
That means a sale today can still be credited to an ad clicked three months ago. Judge a campaign after two weeks and you are reading an unfinished story.
Organic channels stretch this out further still, because SEO work often takes a quarter or more before the first enquiry arrives, and then keeps paying long after the invoice is settled.
Match your patience to your sales cycle. A saree shop in Maninagar sees the effect within days. An interior design firm off S.G. Highway may wait two months while a family argues about the budget. Neither is doing anything wrong.
If your channel mix is still unsettled, our comparison of Google Ads and Meta Ads for lead generation covers how the timelines differ between the two.
What Should You Stop Tracking?
Anything you cannot connect to money. Impressions, reach, follower count and likes all feel like progress, and they mostly measure how much you posted.
They are not useless. Reach confirms a campaign was actually delivered, and followers matter a great deal if you sell through the account itself. The trouble starts when they sit at the top of a report and the cost per customer never appears at all.
A quick test for any number on your dashboard: if it doubled tomorrow, would you make more money? If the honest answer is no, move it to the bottom of the page.
We wrote about the wider version of this problem in the most common digital marketing mistakes small businesses make, and reporting on the wrong things sits near the top of that list.
How Do You Measure ROI If You Sell on the Phone?
This is where most Indian businesses actually live, and where standard advice falls apart. The enquiry arrives on WhatsApp, the deal closes in a showroom, and no tool sees any of it.
Four things close that gap, none of them expensive:
- Ask every caller how they found you. Write the answer in the same place every time, because a crude record kept consistently beats a clever one kept sometimes.
- Use a separate number for ads. Any call to that number came from that campaign, which is cheap to set up and very hard to argue with.
- Track calls properly. Google Ads supports call conversion actions, and lets you import call conversions if you already record outcomes elsewhere.
- Send closed deals back. When a lead becomes a sale weeks later, upload that back into the ad account so it learns what a good lead looks like.
The Google Ads side of this takes a little setup. The asking-every-caller part takes a notebook, and it is the step most businesses skip.
Across the accounts we handle, the businesses that measure well are rarely the ones paying for the best software. They are the ones who ask the same question on every call and write the answer in the same column.
An Example of the Maths
This is an illustration, not a real client’s reported figures. The numbers are round on purpose, so the method stays visible.
A furniture showroom spends ₹1,00,000 in a month. That covers ad budget, agency fee and design time. It gets 200 enquiries, so the cost per lead is ₹500.
Twenty of those enquiries buy something. Cost to acquire a customer is ₹5,000. Average order value is ₹40,000, so the month brought in ₹8,00,000.
ROI is (8,00,000 minus 1,00,000) divided by 1,00,000, which is 700%.
Now the part that changes the decision. If gross margin is 30%, the real gross profit is ₹2,40,000. Against ₹1,00,000 spent, the return is 140%. Still healthy, and exactly one fifth of what the first figure suggested.
Same campaign, two very different answers. The second one is the one worth acting on.
Where Should You Start?
Do these four in order, and do not move on until the one before it is genuinely working.
- Decide what a conversion is. A form, a call over thirty seconds, a WhatsApp message. Write the definition down.
- Count your full spend, including fees and staff hours.
- Track one number for a whole quarter without changing the definition. Cost per customer is the best place to begin.
- Add margin once the rest is steady, so you are measuring profit and not revenue.
Most businesses can do the first three in an afternoon. If you want a second pair of eyes, talk to our team and bring last quarter’s numbers, however messy. Reading them honestly is usually worth more than another campaign.
Frequently Asked Questions
What is a good ROI for digital marketing?
There is no universal figure, because it depends on your margins. A jeweller working on 15% margin needs a far higher return than a software firm on 80%. Compare your ROI against your own last quarter, not a number from an article.
Is ROAS or ROI more useful?
Both, for different jobs. Use ROAS weekly to decide which ads to keep running. Use ROI quarterly to decide whether the whole effort is worth continuing. Reporting ROAS alone tends to hide the fee and salary costs sitting underneath it.
How do I track ROI without any paid tools?
A spreadsheet and discipline will carry most small businesses a long way. Record spend, leads, customers and revenue every month in the same columns. Google Analytics and Google Ads reporting are free and cover the online half.
Why do Meta and Google both claim the same sale?
Each platform reports conversions it believes it influenced, and a customer often touches both. That is why platform totals usually add up to more sales than you actually made. Treat Analytics as your referee and the platforms as claimants.
How long before I can judge whether marketing is working?
Give it at least one full sales cycle, and remember Analytics may credit a source for up to 90 days. For most Ahmedabad businesses a quarter is the shortest honest window.
Should I count organic traffic in ROI?
Yes, though separately. SEO and social both cost money in time or fees, so they belong in the spend column. Keeping them apart from paid stops one channel quietly taking credit for the other’s work.
