FIELD NOTE 01 INTERACTIVE ADS
An Ad That Listens.
What if a car ad asked what you actually needed before showing you what to buy?
Most car ads start with the car: a polished exterior, an empty road, a list of features. But two people watching the same ad might need completely different things. One wants a compact car for city parking. The other needs space for a family and long weekend trips.
I’ve been thinking about what an ad could do with a few useful questions. Instead of sending everyone to the same landing page, it could help someone narrow down their options inside the ad itself.
A short conversation, inside the ad.
Imagine an ad that opens with “What will you mainly use your next car for?” You choose city commuting. It asks whether you prefer a sedan or an SUV, then asks for your budget range and whether you would consider an electric car.
The next screen reflects those answers: two compact options within your budget, with a short explanation of why each might suit you. If you are open to electric, the ad could ask about access to charging before recommending an electric model. If the inventory has no suitable match, it should say so.
The recommendation could be generated from the conversation, but vehicle details, prices, and availability should come from a current product catalog. A helpful explanation is only useful if the underlying facts are right.
Make every question earn its place.
I’d keep the first version short. Ask only questions that change the recommendation, offer a way to skip, and let people browse without providing contact details. A test-drive invitation could come after the recommendation, when there is a clear reason to take the next step.
The trade-off is friction. More answers might improve relevance, but a long conversation could lose someone who just wanted a quick look. The goal would be to make choosing easier.
EXPERIMENT PLAN CONVERSATIONAL ADS
How I’d test it.
This is a hypothetical experiment, not a campaign I’ve run. I’d compare a standard car ad with a short conversational version using the same inventory, offer, comparable audiences, and equal budgets.
Start with three useful questions
Ask about intended use, budget, and vehicle preference. Track where people leave and whether they reach a relevant recommendation. Test an optional electric-car question only when it helps narrow the choice.
Measure what happens after the click
Use cost per attended test drive as the main outcome, with a fixed attribution window. Track bookings and attendance separately: a cheap lead is less useful if the person never turns up.
Check whether the conversation helped
Review completion rates, recommendation clicks, and feedback from the sales team. Repeat with enough traffic before treating an apparent improvement as reliable.
Illustrative numbers only: With $500 spent on each version, 10 attended test drives from the standard ad would cost $50 each. If the conversational version produced 15, the cost would be about $33.33 each. That would be a reason to investigate further, rather than proof that the format always works.
The question I’d keep asking.
Did the interaction help someone make a better decision? If it only added taps, I’d simplify it. If it helped people find a suitable car and arrive at a test drive with clearer expectations, that would make the conversation worth having.