Conversion tracking connects a sale back to the specific ad, keyword, or campaign that caused it, so you stop paying for clicks and start paying for customers.
Why don’t clicks tell me which ads made sales?
A click only tells you someone arrived on your site. It says nothing about whether they bought, called, or filled a form, so two campaigns with identical clicks can produce wildly different revenue.
Every ad platform will happily report clicks, impressions, and cost-per-click, because those are things the platform can see on its own. What it cannot see, without your help, is what happened after the person left the ad and landed on your website. That gap is where most ad budgets quietly leak.
We see this constantly with business owners who come to us frustrated that they are ‘spending on ads but not seeing results’. When we look at their account, the ads are getting clicks at a reasonable cost. The problem is nobody ever told the platform what a result looks like, so it has been optimising for the cheapest clicks rather than the clicks that turn into money.
A cheap click is not a cheap customer. A campaign can deliver a hundred clicks at a low cost and produce zero sales, while another delivers thirty more expensive clicks that produce ten orders. Without conversion tracking, the first campaign looks like the winner and you pour more budget into the thing that is losing you money.
This is the whole reason conversion tracking exists: to draw a line from the ad spend on one end to the actual outcome on the other, so you can compare campaigns on customers rather than on traffic.
What is the difference between a pixel and an event?
A pixel (or tag) is a small piece of code that loads on your website and lets an ad platform recognise visitors. An event is a specific action that pixel reports back, such as a purchase, a form submission, or a phone call.
Think of the pixel as the sensor and the event as the reading. The pixel is installed once, across your whole site, and its job is to sit there watching. It does not know or care what any individual page is for. Google calls its version a tag; Meta calls its version the Meta Pixel; the concept is the same.
An event is what you configure on top of the pixel to say ‘this particular thing just happened, and it matters’. Someone reaching the thank-you page after checkout is a purchase event. Someone submitting a contact form is a lead event. Someone starting checkout but not finishing is an initiate-checkout event, which is what remarketing later chases.
Each event can also carry values. A purchase event should pass the order amount and currency, so the platform can report not just how many sales came from an ad but how much revenue, and calculate a return on ad spend. If the event fires without a value, you know you got a sale but not whether it was a small one or a large one.
Getting these two layers right is the difference between data you can act on and a dashboard full of numbers that mean nothing. The pixel installed with no events configured is the single most common half-finished setup we inherit from clients.
How does attribution decide which ad gets the credit?
Attribution is the rule that assigns credit for a conversion when a customer interacted with more than one ad before buying. Different models credit different touchpoints, and the model you choose changes which campaigns appear to be working.
Real customers rarely click one ad and buy immediately. Someone might see a Facebook ad, click a Google search ad a week later, and finally buy after a remarketing ad. Three ads touched that one sale. Attribution decides how the credit is split between them.
This matters because the model quietly shapes your decisions. Last-click attribution gives all the credit to the final ad before the purchase, which tends to over-reward remarketing and search terms for your brand name, while starving the earlier ads that introduced the customer in the first place. Switch models and a campaign you were about to cut can suddenly look essential.
There is no single correct model. The point is to pick one, understand what it favours, and stay consistent so that your month-to-month comparisons are actually comparing like with like. Switching models mid-quarter and then wondering why the numbers moved is a mistake we have watched businesses make.
A practical note for smaller advertisers: if you run only one channel and a short buying cycle, attribution barely matters and last-click is fine. The more channels you run and the longer people take to decide, the more the model choice distorts your view.
| Model | How credit is assigned | Tends to over-reward | Best for |
|---|---|---|---|
| Last click | All credit to the final ad before purchase | Remarketing and brand-name search | Simple, single-channel setups |
| First click | All credit to the first ad the customer touched | Awareness and discovery ads | Judging what introduces new customers |
| Linear | Credit split equally across every touchpoint | Nothing in particular; treats all equally | Getting a balanced first view |
| Data-driven | Platform’s algorithm splits credit by observed impact | Whatever actually influences buyers | Accounts with enough conversion volume |
A clickSomeone visited your siteCosts you money immediatelySays nothing about intentReported by the platform automaticallyA conversionSomeone did what you wantedTies spend to an outcomeCan carry a revenue valueOnly exists if you set it up
How do I actually set up conversion tracking?
Install the platform’s tag across your whole site, define the events that matter to your business, place each event on the exact action it represents, and then verify every one with a real test before trusting the data.
The cleanest way to install tags is through Google Tag Manager, a free container that sits on your site once and lets you add and manage tags without editing site code every time. You install Tag Manager, then deploy the Google Ads tag, the Meta Pixel, and Google Analytics inside it. This keeps your tags in one place and stops your site turning into a tangle of pasted scripts.
Next, decide what counts as a conversion. For an e-commerce store it is the purchase, plus supporting events like add-to-cart and initiate-checkout. For a service business with no online checkout it is the form submission, the phone call, and the WhatsApp click. Do not track everything; track the handful of actions that represent real business.
Then place each event on the precise trigger, not the page. A purchase event should fire when the order is confirmed, not when the checkout page loads, or you will count everyone who abandoned. A form event should fire on successful submission, not on the button click, or you will count failed submissions.
Finally, pass values where you can. On the purchase event, send the order total and currency so the platform can report revenue and return on ad spend. On a lead event, you can assign an estimated value if a lead is reliably worth a known amount to you.
This is also where connecting your store platform helps. Shopify and WooCommerce both have supported integrations that pass purchase data cleanly, which removes a lot of the manual event work and the errors that come with it.
1Install a tag containerGoogle Tag Manager across the whole site2Deploy platform tagsGoogle Ads, Meta Pixel, Analytics3Define your key eventsPurchase, lead, call — not everything4Trigger on the actionFire on the event, not the page load5Pass a valueOrder total and currency where possible6Test with a real transactionConfirm each event actually fires
Which events should I actually track for my business?
Track the handful of actions that represent real outcomes for your business, plus the earlier steps that feed remarketing. For a shop that means purchase and add-to-cart; for a service it means form, call, and message clicks.
Tracking too much is as unhelpful as tracking nothing. A dashboard cluttered with page views, scroll depth, and clicks on your logo buries the two or three numbers that decide where your budget goes. Start with the money-making action and work backwards.
For an online store, the primary event is the purchase, carrying its value. Add-to-cart and initiate-checkout are secondary events worth tracking because they tell you where people drop off and give remarketing something to chase. For a lead-generation business, the primary events are form submissions, phone calls, and WhatsApp or messenger clicks, which in Pakistan often carry as many enquiries as a form does.
Match the event to how customers really reach you. A local service business in Lahore or Karachi may take most of its enquiries by phone or WhatsApp rather than through a form, so a setup that only tracks form fills will report almost none of the real demand and make good ads look like failures.
Once the primary events are solid, you can layer softer signals like newsletter sign-ups or brochure downloads. But get the sale or the lead working first and verified before adding anything else.
Why does cash-on-delivery break conversion tracking in Pakistan?
Cash-on-delivery records a sale when the order is placed online, but the actual payment only happens later at the door, and a share of those orders are refused or returned. So the purchase event over-counts revenue unless you reconcile it against delivered orders.
Most conversion tracking was designed around online payment: the customer pays at checkout, the purchase event fires, and the sale is real. Cash-on-delivery, which still dominates a large share of Pakistani e-commerce, does not work that way. The order is placed online but the money changes hands days later when the courier arrives, and not every order survives that gap.
This creates a specific problem. Your purchase event fires at order placement, so the platform reports the full order value as revenue and optimises towards it. But if a meaningful percentage of COD orders are refused at the door, returned, or turn out to be fake, your reported return on ad spend is inflated and the platform is learning from orders that never became cash.
The practical fix is to track the order-placed event as your live signal, because you need something to fire in real time for the platform to optimise on, and then reconcile against delivered-and-paid orders periodically. Some businesses use offline conversion imports to feed the confirmed, paid orders back into the ad platform so it eventually optimises for orders that actually complete, not just orders placed.
This is exactly the kind of local reality that generic tracking guides ignore, and it is why we publish practical content for Pakistani SMEs on COD and payments. A tracking setup that works in a card-first market can quietly mislead you here if nobody adjusts it for how people actually pay.
What goes wrong with conversion tracking?
The common failures are tracking nothing at all, double-counting the same conversion, firing events on page load instead of the real action, and letting duplicate tags or broken thank-you pages silently stop the data.
The first and biggest failure is running ads with no tracking configured, so the account optimises for clicks and you have no idea which spend produced sales. It is astonishingly common and it is the reason many businesses believe ‘ads don’t work’ when in fact their ads were never pointed at a goal.
Double-counting is the next trap. If the same purchase event is installed both in Tag Manager and hard-coded on the site, it fires twice and every sale is counted as two. Your return on ad spend looks fantastic and it is fiction. This usually happens when a new setup is added without removing the old one.
Firing on the wrong trigger quietly poisons the data. An event placed on the checkout page load counts abandoners as buyers; an event on a button click counts failed form submissions as leads. The numbers look plausible, which is what makes this one dangerous — nobody questions a figure that seems reasonable.
Then there are the silent breakages: a redesigned site drops the tag, a changed thank-you page URL means the event stops firing, or a consent banner blocks the pixel until the user accepts. Any of these can flatline your conversions overnight while spend continues. This is why tracking is not a set-and-forget job; it needs checking after every site change.
How do I check my tracking is actually working?
Complete a real conversion yourself and confirm the event registers, use the platform’s built-in diagnostic tools to catch errors and duplicates, and re-verify after any change to your website.
The only test that fully counts is a real one. Place an actual order or submit an actual form, then check that exactly one conversion appears with the correct value. Doing this catches the errors that no amount of looking at code will reveal, because it tests the whole chain from click to reported conversion.
Use the free diagnostic tools each platform provides. Google Tag Assistant and the Tag Manager preview mode show you which tags fired on which page as you click through your own site. Meta’s Events Manager has a test-events mode that shows events arriving in real time. These tools flag duplicates, missing values, and events that never fire.
Cross-check totals against your own records. Your ad platform’s reported purchases should line up roughly with the orders in your Shopify or WooCommerce admin. A large gap in either direction — many more or many fewer — means something is miscounting, and it is worth finding out which before you make budget decisions on the number.
Build re-checking into your routine. Any time the website is redesigned, a checkout flow changes, a plugin updates, or a thank-you page URL moves, test the events again. We treat tracking verification as part of every site change rather than a one-time task, because a setup that was perfect last month can be silently broken today.
Common questions
Do I need conversion tracking if I only run a small budget?
Yes, and arguably more so. A small budget has no room to waste on clicks that never convert. Tracking even a single event — the purchase or the lead — lets the platform steer your limited spend towards the actions that make money instead of the cheapest traffic.
Is Google Analytics the same as conversion tracking?
Not quite. Google Analytics measures site behaviour and can record conversions, but to optimise ad campaigns the conversion signal needs to reach the ad platform itself, either through the native Google Ads tag or by importing conversions from Analytics. Analytics alone will show you what happened but won’t automatically teach your ads to chase it.
How long before conversion data becomes reliable?
It depends on volume. The platform needs a steady flow of conversions before its optimisation and any data-driven attribution become trustworthy — a store getting many orders a day builds a usable picture in weeks, while a low-volume advertiser may need longer. Until then, read the numbers as a direction, not a verdict.
Will iOS privacy changes and cookie limits break my tracking?
They reduce it rather than break it. Browser restrictions and app privacy prompts mean some conversions go unattributed, so reported numbers can undercount real sales. Server-side tracking and the platforms’ modelled conversions help close the gap, but you should expect your dashboards to show slightly fewer conversions than actually occurred.
Can I track phone calls and WhatsApp messages as conversions?
Yes, and in Pakistan you often should. Google Ads can track calls from ads and from a website number, and clicks on a WhatsApp or messenger link can be set up as events. For many local service businesses these carry more real enquiries than web forms, so leaving them untracked hides your best-performing ads.
Who should install this — me or a developer?
Basic setups through Google Tag Manager and a store platform’s native integration can be done without a developer. But triggering events on the right action, passing correct values, handling COD reconciliation, and verifying everything is where mistakes hide, and those are worth a specialist. A wrong setup is worse than none because it gives you confident, false numbers.
