Most marketing dashboards are crowded with numbers and settle nothing. The marketing metrics that matter are the small set that forces a decision: spend more here, stop that, fix the follow-up before Monday. Everything else is a diagnostic, useful only when one of those decisions goes wrong.
We build reporting for the person who signs the cheques, so we keep it blunt. If you run search, ads and content as one programme, six numbers will carry almost every call you make this year. This post sets out which six, why the popular numbers do not qualify, and how a 15-person company keeps a report alive past month three.
Marketing metrics that matter change a decision, not a mood
Here is the test we apply to every line in a report. Name the number, then name the action you would take if it moved sharply in either direction. If no action follows, it is a diagnostic and belongs in a second tab. If an action follows, it belongs on the front page with a target beside it.
This matters more in a small company than a large one. A 200-person marketing team can afford analysts who read dashboards for a living. You have a founder, maybe a marketing manager, and a Friday afternoon. Reporting that costs more attention than it returns gets abandoned before long, and then you are back to guessing.
What counts as a vanity metric?
A vanity metric moves without telling you what to do next. Impressions, followers, likes and raw sessions all describe activity, not outcomes. They are useful as diagnostics when a decision metric drops, because they help explain why. They fail as goals because you can grow them while revenue stays flat.
Impressions, followers and sessions are not lies. They are inputs. When qualified enquiries fall and impressions fell with them, you have a reach problem. When qualified enquiries fall while impressions held steady, you have a message or landing page problem, and that is a different fix.
Vanity versus decision metrics, side by side
Use this as a swap list. Each number on the left has a replacement on the right that answers the question you actually had in mind when you asked for the report.
| Vanity metric | Decision metric to use instead | Question it answers |
|---|---|---|
| Impressions and reach | Qualified enquiries by source | Which channel produced people worth calling? |
| Followers and likes | Cost per qualified enquiry | What does one useful conversation cost here? |
| Sessions and pageviews | Conversion rate by pipeline stage | Where do buyers stop moving? |
| Average keyword position | Enquiries from organic search | Is search bringing buyers or browsers? |
| Cost per click | Acquisition cost against customer value | Can we afford to keep buying this traffic? |
The six numbers a growth team runs on
Six is not a magic count. It is roughly what a small team can update honestly every month without hiring someone to do it.
- Qualified enquiries by source. Not total leads. Count only the people who match your buyer and asked something real, split by Google search, ads, WhatsApp, referrals, JustDial or IndiaMART.
- Cost per qualified enquiry. Channel spend plus the agency or salary cost attached to it, divided by the qualified enquiries that channel produced.
- First-response time. Minutes from enquiry to a human reply. It usually needs no budget to improve, and it is easy to overlook.
- Conversion rate by pipeline stage. Enquiry to conversation, conversation to quote, quote to order. One rate for the whole funnel hides where the leak sits.
- Acquisition cost against customer value. What you paid to win a customer, next to what that customer is worth across the relationship, not on the first invoice.
- Pipeline velocity. How long a deal takes from first enquiry to signature, and how much value is sitting in each stage right now.
First-response time deserves attention it rarely gets. An enquiry that waits until tomorrow is competing with suppliers who replied while the buyer was still on their phone. Fix nothing else this quarter and fix this, because a tidy lead handling setup usually pays back faster than more ad spend.
Cost per qualified enquiry is where a stretched budget often gets rescued. It exposes the channel that looks cheap on clicks and expensive on customers. If yours is climbing, work through the levers that pull a cost per lead down before you cut spend, because the fix is often the offer or the form.
Acquisition cost only means something when it sits next to customer value. A clinic, a manufacturer and a property consultant can all afford very different acquisition costs, because their repeat value differs. Get the value side roughly right, then let it set your ceiling and your split across channels.
How do you track qualified enquiries by source without expensive tools?
Ask every enquiry where it came from, and record the answer in one sheet before anything else. Add call tracking numbers only if call volume justifies the cost. Tag your forms and WhatsApp links so the source arrives with the lead. Consistent manual tagging beats an unused analytics setup every time.
- Use one form with a hidden source field, not five forms nobody maps back to a channel.
- Give each channel its own WhatsApp link or landing page so the origin is unambiguous.
- Ask on the call: "how did you hear about us?" Record the answer the same day.
- Keep one sheet or CRM as the only place a lead is logged. Two places means no places.
- Define "qualified" in writing, so the same lead is counted the same way in June and in December.
- Review sources monthly and merge the ones too small to act on.
The definition of qualified is the part teams skip. Write it down: budget band, location, service needed, and whether the person asked a specific question. Our own intake uses bands of under Rs 50,000, Rs 50,000 to 2 lakh, Rs 2 to 5 lakh, and Rs 5 lakh plus, because a band is easier to answer honestly than an exact figure.
What should a one-page monthly report contain?
One page, six rows, three columns: this month, last month, and the decision you took. Add a two-line note on what changed and what you will try next. If it takes more than an hour to assemble, it will not survive a busy quarter, so cut rows until it does.
- Pull the six numbers on the first working day of the month, the same day every month.
- Write last month's decision and whether it worked, in one line.
- Mark one number as this month's focus. Only one.
- List the experiments running and the date each one ends.
- Send the page to everyone who can act on it, sales included.
Keep the diagnostics in a second tab: impressions, average position, click-through rate, page speed, bounce. You will want them the month something drops, and you will not want them in the meeting. If the monthly page keeps growing, the team is reporting to look busy rather than to decide.
We run this cadence with clients through weekly checkpoints and a single monthly page, and we reply to questions within one business day. If your current report takes half a day to build and still leaves you unsure what to change, tell us what you are measuring now and we will show you the shorter version.
Where these numbers mislead you
Attribution is imperfect and always will be. A buyer sees your ad on Monday, searches your brand name on Thursday, and tells you they found you on Google. Treat source data as directional, read trends across quarters, and never rebuild a whole budget on one noisy month.
Small samples are the second trap. A handful of enquiries cannot produce a reliable conversion rate, so two lost deals look like a collapse. For low-volume, high-value businesses, watch pipeline value and first-response time closely, and read conversion rates over a quarter instead of a month.
Lag is the third. Organic search compounds slowly, so judging work on search visibility by a single month of enquiries will make you cancel the thing that was about to work. Paid campaigns report faster, which is why teams over-trust them, and the honest way to compare Google Ads with Meta Ads is cost per qualified enquiry read over months.
Watch what the numbers do to behaviour, too. Pay a team on lead volume and you will get volume, most of it unqualified. Track qualified enquiries and cost per qualified enquiry together, and the incentive points at the outcome you wanted. We judge our own paid campaign work the same way, because clicks are easy to buy and customers are not.
Frequently asked questions
How often should we review these numbers?
Monthly for the full set, weekly for first-response time and enquiry volume. Anything faster invites overreaction to noise. Anything slower and you can lose a quarter before you notice a channel has stopped working. Weekly checkpoints with a monthly decision page is a rhythm most small teams can hold.
Do we need a paid analytics tool to start?
No. A spreadsheet, tagged forms and one source question on every call will carry you a long way. Buy a CRM when the sheet starts breaking, not before. Tool pricing changes often, so check the vendor's own page rather than a comparison article when you do start shopping.
What is a reasonable cost per qualified enquiry in India?
It depends on your ticket size, not on a benchmark. A service worth a few thousand rupees cannot carry the same acquisition cost as a project in the Rs 2 to 5 lakh band. Set your ceiling from customer value and margin, then compare this month against your own last three.
Who should own the monthly report in a 15-person company?
One named person, and preferably not the founder. Give it to whoever sits closest to the enquiries, usually the marketing manager or whoever handles the sales inbox. The founder reads it and picks the focus number. Shared ownership means it quietly stops happening.
Pick one number and fix it this month
Start with first-response time, because it needs no budget and no new tool. Then add qualified enquiries by source, then cost per qualified enquiry. Three months in, you will have a report that tells you where the next rupee should go, which is the only job a report has.
We work in that order with clients: Connect to understand the business, Build the tracking and the pages, Launch, then Optimise on real numbers, with campaigns typically live in two to four weeks. If you want a second pair of eyes on what you are measuring, send us your current report and we will tell you which rows to delete.
