It is not enough for a company’s head to know how many impressions and clicks there were. What matters is what each customer cost, whether the investment paid off and how much a buyer will bring in over the whole relationship. When marketing reports in these terms, the conversation with the business becomes concrete. Below is a reference guide to the key performance indicators for advertising, two worked examples and practical checklists for the Tajik market.
Formulas worth keeping to hand
| Metric | How it is calculated | What it shows |
|---|---|---|
| CR (Conversion Rate) | Conversions / Sessions × 100% | the share of visits that ended in a target action |
| CPC (Cost per Click) | Spend / Clicks | the price of one click |
| CPA / CPL | Spend / Conversions | the cost of an order or a lead |
| CAC (Customer Acquisition Cost) | Marketing and sales costs / New paying customers | the price of acquiring a buyer |
| AOV (Average Order Value) | Revenue / Number of orders | the average order value |
| ROAS | Revenue from advertising / Advertising spend | how much revenue each somoni in media brings in |
| ROI | (Revenue − Costs) / Costs | the real return, taking all costs into account |
When calculating ROI, include everything in costs: not just media buying but also creative production, platform commissions, discounts and returns. Otherwise the return will look better than it is.
Customer value over the whole period
In simplified form, LTV = average order value × number of purchases in the period × margin. Subscription models use a different version: LTV = ARPU / Churn, or ARPU × average customer lifetime.
Where the profitability line lies
Break-even CPA (or CAC) = contribution margin from a customer over their lifetime × the share you are willing to spend on acquisition. As long as the actual acquisition cost does not exceed this threshold, advertising is not running at a loss.
Two worked examples
Leads for a service
Say 10,000 somoni was spent on a campaign. It brought 4,000 clicks, so CPC = 10,000 / 4,000 = 2.5 TJS. There were 160 leads, so CPL = 10,000 / 160 = 62.5 TJS. A quarter of those who enquired became buyers — 40 people — and CAC = 10,000 / 40 = 250 TJS.
Suppose the contribution margin per order is 320 TJS. Acquisition costs less (250 < 320), so the campaign is profitable. A rough ROI estimate: (40 × 320 − 10,000) / 10,000 = (12,800 − 10,000) / 10,000 = +28%.
A shop with repeat purchases
- An average order of 250 TJS with a 40% margin gives 100 TJS of income per order.
- Over six months, a customer buys twice on average: 100 × 2 = 200 TJS for the period.
- The company is willing to spend 60% of this amount on acquisition, so the maximum CAC is 120 TJS.
With an actual CAC of around 110 TJS, the campaign can be scaled up. If it has reached 150 TJS, it is time to lower the cost per click, improve the website’s conversion rate or encourage repeat orders.
Six rules for reading reports
- The same attribution window. For performance campaigns it is usually shorter, for media campaigns longer. Results calculated with different windows cannot be compared.
- Branded traffic separately. Searches containing the company name improve the averages. Show branded and non-branded campaigns in different columns.
- The whole funnel. Cheap clicks but expensive leads? The weak spot is the landing page or the offer itself.
- Segments. Region, device, new and returning users. Averages hide both failures and growth opportunities.
- Reconciliation with the CRM. A lead from advertising is not yet a qualified lead. Keep two levels in the report: MQL and SQL or sale.
- Notes on changes. Changed the price, removed a discount, moved a form — mark the date on the dashboard, otherwise your conclusions will be skewed.
Mistakes that make the numbers lie
- Passing ROAS off as ROI. The former does not account for production, discounts and delivery, so a nice-looking report can coexist with a shortage of cash in the bank.
- Counting a lead as a customer. Cheap leads with poor qualification turn into a loss-making CAC.
- Not deducting returns and cancellations. Revenue in reports needs to be adjusted.
- Judging all channels by the same yardstick. Search works with existing demand, video builds reach — and their KPIs should differ.
- Looking only at averages. In a particular region the cost per lead can be noticeably higher, and the average will hide it.
- A report without conclusions. Every report should end with two or three hypotheses and a testing plan.
A one-page report: what it should include
- the campaign goal (leads, sales or awareness) and the KPIs that match it;
- the attribution window and notes on changes;
- branded and non-branded campaigns shown separately;
- the chain: CPC → CTR → click-to-lead conversion → lead-to-purchase conversion → CAC;
- unit economics: average order value, margin, LTV, break-even point;
- a few hypotheses for the next period and the expected effect of each.
Questions and answers
Which matters more — ROAS or ROI?
You need both. ROAS shows how efficiently the traffic was bought; ROI shows how much the business actually earned after all costs.
How can you quickly tell whether a campaign is working?
In the first few days, go by early signals: CTR, CPC, add-to-carts, first leads. If leads are coming in but sales are few, the cause most likely lies with the sales managers, logistics or price.
When do you need LTV?
When customers come back or pay by subscription. For one-off purchases, assess the income from the first deal and look for upselling opportunities.
If you want to build a KPI and reporting system tailored to your funnel, write to us. Our digital team handles performance campaign set-up and creative testing, and you will find an overview of all SKY Central Asia’s online tools in the digital section.
Leave a comment