Lead Source Tracking: Which Ad Sold the Car?
Most car dealers know what a Facebook lead costs but not which ad sold last month's cars. Where the source gets lost, and how to carry it to delivery.
Ask a car dealer what a Facebook lead costs and you get a number in seconds. Ask which ad sold the 22 cars delivered last month and the room goes quiet. Car dealer lead source tracking usually stops at the enquiry: the ad platforms count leads, the sales report counts cars, and no field connects the two.
The position this post takes: judge every channel on what it costs per delivered car, counted by the month the enquiry came in, and stop judging it on cost per lead. That only works if the lead's source survives three handoffs most showrooms never look at. Those are the move to a salesperson's own WhatsApp, the booking form, and the loan and delivery desk.
A car dealer's lead source is rarely lost at the ad. It is lost later, when the chat moves to a salesperson's personal phone, when the booking form is written without a source field, and when the loan and delivery team works from a new file. Keep one record from first enquiry to delivered car, set the source at the first message instead of asking at handover, and divide each channel's spend by the cars its enquiries delivered, counted by enquiry month. The cheapest lead is often not the cheapest car, and "walk-in" is usually a step in the journey, not a source.
Where does a car dealer lose the lead source?
A car dealer loses the lead source after the enquiry, not at the ad. Every platform records where its leads came from. The record breaks at three handoffs inside the showroom, and by delivery the car has no source at all, or the wrong one.
Picture a used-car dealer in Kepong with six salespeople, around 80 cars in stock, and an ad budget spread across Facebook lead forms, TikTok, listing packages on Mudah and Carlist, and a small Google search campaign. Each platform sends its own report. None of them knows which buyer drove away in which car.
Handoff 1: the chat moves to a personal phone
The enquiry lands on the company number or in a lead form. Within the hour, a salesperson replies "I WhatsApp you from my number ya", and the conversation moves to a phone the business cannot see. The source stays behind with the first message. The new chat starts with no history, no ad name, nothing.
Handoff 2: the booking form
Two weeks later the buyer pays a booking fee. A typical booking form asks for the name, IC number, the car, the agreed price, the deposit and the salesperson's name. It rarely asks where the buyer first came from, because the form was built for the bank and the ownership transfer, not for marketing.
Handoff 3: the loan and delivery desk
The file then moves to whoever handles the loan submission, the Puspakom inspection and the JPJ transfer. That person never saw the enquiry. It is also the stage where deals stall for weeks, as we covered in why a two-day loan approval still turns into a weeks-long sale. By delivery the record is a sales file, the month-end report lists cars by salesperson and model, and the source column, if there is one, says "walk-in", "referral" or nothing.
So the month-end meeting puts two reports side by side that share no field. Facebook says 300 leads at RM20 each. The sales report says 22 cars. Nobody can say how many of the 22 came from the 300, and next month's budget gets set on the number that is easiest to read: cost per lead.
Is a walk-in really a lead source?
Usually not. For most car buyers, walking in is the last step of a journey that started on a screen, so "walk-in" on a sales report mostly means "we did not record where they came from".
The Malaysian data on this is old but clear. A Google and TNS study found that 86% of Malaysian car buyers researched online before buying, 74% said they were influenced by the ads they saw, and buyers expected to visit only 2.5 dealerships before deciding, down from 3.5 in 2013 (Google and TNS, reported by Paul Tan, April 2017). Newer work elsewhere in Asia points the same way. A Google and Kantar study of 1,022 recent car buyers in India, published in September 2026, counted an average of 6.9 touchpoints on the way to a purchase (Autocar Professional, September 2026). Different country, same shape: several screens first, then a showroom.
The standard fix, given on dealer forums since at least 2013, is to ask the customer at the time of sale (DealerRefresh forum). That is the least reliable moment to ask. A buyer who saw your TikTok in August, asked about a listing on Mudah in September and walked in during October will answer with whatever is freshest, or with "just passing by". The answer at handover rewards the last touch and the channel with the catchiest memory. It is a guess written down as data.
The guess has a cost. A dealer who reads "40% walk-in" on the sales report and trims the Facebook budget to pay for a showroom event may be cutting the very thing that produced the walk-ins. The fix is a rule, not a question: a lead counts as a walk-in only when the buyer's phone number has no earlier record. Search the number before you create the lead.
Why is cost per lead the wrong number for car dealers?
Because a car is sold three filters after the lead: the viewing or test drive, the booking, and the loan. A channel can win on cost per lead and lose at every filter after it, and cost per lead cannot see any of them.
The loan filter alone moves serious money. Hire-purchase approvals in Malaysia reached a record RM10.7 billion in December 2025, against car loan applications worth RM18 billion in the same month (DOSM data reported by Malay Mail, April 2026). The two figures do not map one to one, since an approval can land weeks after its application. The point for a dealer is simpler: the bank decides between booking and delivery, and the loan outcome belongs on the same record as the source.
Here is the arithmetic with round, illustrative numbers for the Kepong dealer. They are not benchmarks. They show how the order can flip once you count cars instead of leads.
| Channel | Monthly spend | Enquiries | Cost per lead | Cars delivered | Cost per delivered car |
|---|---|---|---|---|---|
| Facebook lead form | RM6,000 | 300 | RM20 | 4 | RM1,500 |
| TikTok | RM3,000 | 100 | RM30 | 3 | RM1,000 |
| Mudah and Carlist listings | RM2,500 | 50 | RM50 | 5 | RM500 |
| Google search | RM2,000 | 25 | RM80 | 3 | RM667 |
On cost per lead, the Facebook form is the star and Google is the problem. On cost per delivered car, the order turns over. The reason is not that one platform is good and another bad. A one-tap lead form brings volume, including people who were curious, not ready yet, or not able to get financing. A buyer who messages about a specific listing has already picked a model, a year and a price band. Fewer of them arrive, and in this example more of them make it through the loan.
There is a second reason to know the source, and it lands on the salesperson rather than the budget. Across the MSME sales teams we work with, organic enquiries tend to explain what they need in the first few messages, while ad leads usually need more of a push. When the source is visible on the first message, the salesperson knows which first reply to send: a lead-form lead needs a question before a price list.
Why does the month you count matter?
Because the buyers who enquire in a month are not the buyers who take delivery in that month. The last Malaysian measure we could find, an iCar Asia study run by Frost & Sullivan, put the average gap between first interest and purchase at 2.4 months (iCar Asia, 2013). Add a loan and, for popular new models, a waiting list, and the gap from enquiry to keys runs to weeks at best and often months.
So compare like with like. September's ad spend is judged against the cars delivered from September's enquiries, whenever those deliveries land. A dealer who divides October's spend by October's deliveries is crediting this month's ads with last quarter's buyers, and will cut a channel in the very month it starts to work.
Frequently Asked Questions
How do you track a lead's source from first enquiry to delivered car?
Keep one record for the buyer from the first message to the keys, and set the source at the first message, not at the end. The six steps below work in a spreadsheet, though they hold up far better in a CRM where nobody has to remember to type them.
How to Track a Car Lead's Source From Enquiry to Delivery
The lost reason in step 5 is the same field we argued for in why dead car leads have an expiry date. A buyer lost to "loan rejected" in March may be financeable by September, and the source on that record tells you which campaign found them in the first place.
What changes in the ad budget once you can see it?
Money moves from the channel with the cheapest leads to the channel with the cheapest cars, and a few decisions that felt obvious reverse.
| Decision | Judged on cost per lead | Judged on cost per delivered car |
|---|---|---|
| Which channel gets more money | The lowest RM per lead | The lowest RM per car delivered |
| A listing platform with few enquiries | Looks expensive, gets cut | Kept if its few enquiries keep turning into cars |
| A month full of walk-ins | Credited to the showroom | Matched to earlier enquiries from the same phone numbers |
| A new campaign after four weeks | Judged on this month's sales | Judged once its enquiry month has had time to deliver |
| High lead volume, few deliveries | Celebrated in the meeting | Checked against loan outcomes and lost reasons first |
| Salesperson performance | Leads handled | Cars delivered, read alongside the source mix each person was given |
The last row matters more than it looks. When cars are counted by source, a salesperson who was handed mostly lead-form leads is no longer compared on raw sales with a colleague who got the referrals. The source mix explains the gap, and the conversation shifts from "why so few" to "which first reply works for this kind of lead".
Two things to avoid once the numbers arrive. Do not cut a channel on one bad month: give it a full cycle from enquiry to delivery. And move money in steps, one channel at a time, so you can see what each change did. Move everything at once and next quarter's numbers will not tell you which change worked.
What can Raion HUB track, and what can't it?
In Raion's automotive CRM, Facebook and Instagram lead ads, TikTok lead ads, website forms, Google Forms and Sheets, and WhatsApp click-to-chat links come in as named lead sources, and keyword rules can sort a WhatsApp chat into the right source from its first message. The source fills itself on the lead, so step 2 happens without anyone typing it.
The team replies from the shared business number inside the inbox instead of a personal phone, which closes handoff 1. Pipeline stages are yours to name, so Booked, Loan submitted, Loan approved and Delivered sit on the same record as the source, and every stage or owner change is logged with who made it and when. Filter the table by source and stage, or export the leads to Google Sheets and count delivered cars per source per enquiry month.
Three honest limits. Raion HUB has no integration with Mudah or Carlist: a listing enquiry arrives only when the buyer messages your WhatsApp number, which is why step 1's separate link or code matters. Google Ads is not a direct lead source, so point Google traffic at a website form. And it does not import ad spend, so there is no cost-per-car report inside the app. The spend column comes from your ad managers, and the division is yours.
For the wider enquiry-to-aftersales playbook, see our guide to WhatsApp automation for car dealerships.
The bottom line
Most dealers can price a lead and cannot say which ad sold a car, because the source is dropped at the personal phone, the booking form and the loan desk. Set the source at the first message, keep one record until the keys are handed over, treat "walk-in" as a source only when the number has never been seen, and judge each channel on cost per delivered car by enquiry month. Do that for one full cycle before you touch the budget, and the next budget will be built on cars, not clicks.
