Table of Contents
The stage of the buying process captured by each channel
How targeting works in each one
Ad formats and how they appear
The product feed: differences between the two systems
How performance is measured on each channel
Compared costs: what to expect from each
What kind of leads each channel generates
When to use only VLA, when to use only ChatGPT Ads, and when to use both
The question dealerships ask most: do ChatGPT Ads replace portals?
Dealcar and the two channels
Frequently asked questions
The stage of the buying process captured by each channel
The most important difference between the two channels is not technical: it is the moment they reach the buyer.
Google Vehicle Ads intercepts active, specific search. The buyer has already more or less decided what they want: they search for "second-hand diesel 2020 Volkswagen Golf Barcelona" or "used family SUV under 20000 euros". They already know the type of car, sometimes the specific model, and they are comparing options available now. This is high and specific intent.
Read also how the car search process is changing with ChatGPT in 2026.
ChatGPT Ads intercepts the conversational consideration phase. The buyer is still defining what they need: they explain to ChatGPT that they have two children, drive 50 km a day, do not want petrol, and have an £18,000 budget, and they ask for recommendations. This buyer has not decided on the model yet, but they have a very clear need and are receptive to specific options.
Neither of the two stages is better in the abstract. They are different. VLA captures the buyer closer to the closing of the deal. ChatGPT Ads captures them higher up the funnel, when they can still be influenced by the options they see.
How targeting works in each one
Google VLA works by vehicle model: the advertiser uploads the stock feed and Google shows the cars when someone searches for that specific model or related terms. Targeting is by keywords and search signals that Google automatically interprets.
ChatGPT Ads works through context hints: descriptions in natural language of the buyer's situations and needs. There are no keywords. The advertiser writes something like "person looking for a reliable second-hand family car for intensive use" and OpenAI's system determines in which conversations to show that ad based on the context of the active conversation.
Check what ChatGPT Ads context hints are and how to use them to sell cars.
The practical consequence is that VLA has more predictability: the dealership knows their ads appear when someone searches for that model. ChatGPT Ads has more contextuality: the ad appears in conversations where the context is relevant, even if the buyer has not typed the name of the model.
For a dealership with varied stock, VLA allows targeting by model and managing different bids for high-demand cars. ChatGPT Ads requires defining the buyer profile well, rather than the product.
Ad formats and how they appear
In Google VLA, the ad appears at the top of the search results in a horizontal carousel format, featuring the car's photo, make, model, price, and the dealership's name. The user is on the Google results page, viewing multiple cars side by side.
In ChatGPT Ads, the ad appears below the response generated by the model, in a separate block labelled as sponsored. The user is reading the AI's response to their question, and the ad appears right after. The format includes a vehicle image, title, description, and link to the detail page.
The difference in context matters. On Google, the user expects to see ads interspersed with results: it is part of the familiar format. On ChatGPT, the user is in the middle of a conversation, and the ad arrives just as they have received a relevant response for their decision. This can lead to greater receptivity or more interruption, depending on how the ad is designed.
The product feed: differences between the two systems
Both channels work with a vehicle product feed, but the format and fields are different.
Google VLA has a specific feed format for automotive with sector-specific fields: number of passengers, fuel type, body style, transmission, year of manufacture, mileage, and condition (new or used). Google validates that the feed meets the requirements and processes it to display the cars in the VLA carousel.
ChatGPT Ads uses OpenAI's generic product feed: item_id, title, description, url, image_url, price, availability, seller_name, and seller_url. There are no specific fields for vehicles in the current official format. The title and description are the fields the AI uses to understand what car it is and when to display it.
Read also how to prepare the car product feed for ChatGPT Ads.
This means that the quality of the texts in the feed matters more in ChatGPT Ads than in VLA. In VLA, the structured format of specific fields tells Google exactly what kind of car it is. In ChatGPT Ads, the title and description are the information the system uses to determine contextual relevance.
A good description for a car in the ChatGPT Ads feed is not "Volkswagen Golf 2020 diesel 115HP": it is "Volkswagen Golf 2020 automatic diesel, 67,000 km, serviced at official workshop, ideal for daily city use or long journeys, C-label, transfer included".
How performance is measured on each channel
Google VLA has a mature measurement system integrated with Google Ads and Google Analytics. Conversions are set up in Google Ads, tracking works with the Google pixel, and performance data by vehicle model is available with good granularity.
ChatGPT Ads has its own measurement system in OpenAI Ads Manager: impressions, clicks, CTR, CPC, and conversions with the OpenAI pixel. The channel is new, and historical data for benchmarking is scarce. For traffic reaching the website from ChatGPT Ads without the pixel installed, the oppref=chatgpt parameter in the URL allows it to be identified in Google Analytics.
Attribution is more complex in ChatGPT Ads because the buying cycle can be long: the buyer sees the ad on ChatGPT, visits the website days later from another channel, and buys over the phone. In VLA, the cycle can also be long, but tracking is more mature.
Read also how to measure ChatGPT Ads leads and sales in a dealership.
To compare the actual performance of the two channels, the most useful metric is cost per qualified lead, not CPC. A lead coming from ChatGPT Ads that reaches the detail page of a specific car with contact details can be worth more than a cheap VLA click that bounces in 10 seconds.
Compared costs: what to expect from each
Google VLA does not have a standard public cost because it works by auction. In Spain, the typical CPC for used car dealerships varies by model and area, but is in a range of between 0.30 and 1.50 euros per click on used car searches.
ChatGPT Ads has a recommended bid of 3 to 5 dollars per click according to OpenAI, with a CPM starting from approximately 25 dollars. This cost per click is higher than VLA's, but the audience receiving the ad is in a context of greater receptivity. If the click-to-lead conversion rate is higher in ChatGPT Ads, the cost per lead can be comparable or lower.
In the first experiments published by advertisers in other sectors, the CTR of ChatGPT Ads is between 1.5% and 3.2%, compared to a 0.35% average for conventional display. Specific automotive data for Spain is not available yet as the channel has just opened.
What kind of leads each channel generates
The lead coming from Google VLA already knows what model they are looking for and is comparing available options. It is a lead with high and specific intent: if they reach the detail page of a specific car, they are very likely to be considering that car.
The lead coming from ChatGPT Ads might be in an earlier decision stage. They have described their need to the AI, received a response with recommendations, and clicked on an ad for a car that fits what the AI suggested. It is a contextually qualified lead but with a potentially longer decision cycle.
In practice, this translates into different sales processes. The VLA lead might be ready to visit the car this week. The ChatGPT Ads lead might need more information and more follow-up before committing to a visit.
When to use only VLA, when to use only ChatGPT Ads, and when to use both
For a dealership not using either of the two channels, the recommended order is to start with VLA. The intent is more direct, tracking is more mature, and there is more reference of what works in the sector. Once VLA is optimised and generating leads with a known return, add ChatGPT Ads as a second layer.
Also check the complete Google VLA guide for dealerships.
For a dealership already using VLA with good results and wanting to diversify acquisition channels, ChatGPT Ads is the natural next channel to test. The initial test budget can be 200 to 400 euros per month to gather enough data in 30 days.
For a dealership that has a very specific stock profile (cars of a single segment, very specific geographic area, customer with a defined profile), ChatGPT Ads can be especially efficient because the context hint can describe that buyer very accurately and the feed only contains relevant cars.
Both channels together make sense when the budget allows and when there is capacity to manage leads from two channels with different follow-up processes. It makes no sense to activate ChatGPT Ads if VLA leads are already not managed well due to lack of capacity.
The question dealerships ask most: do ChatGPT Ads replace portals?
No, at least not in the short term. Portals (Coches.net, AutoScout24, Wallapop) remain the channel where most used car buyers in Spain start their search. They have decades of indexed content, millions of vehicle detail pages, and a consolidated audience.
ChatGPT Ads does not compete with portals on the same playing field: it is not a marketplace where the buyer filters by model and price. It is a conversational channel where the ad appears when someone is asking for help to decide. They are different parts of the search process.
What might change in the medium term is the proportion of buyers starting on ChatGPT instead of Google or a portal. If that percentage grows, the dealership that already has a presence on ChatGPT Ads will have an advantage over the one that has not yet tested the channel.
Dealcar and the two channels
Dealcar automatically manages the stock feed for Google VLA through Dealcar Boost: when a car enters the stock, it appears in VLA ads; when it is sold, it disappears. Dealcar is developing the same integration for ChatGPT Ads, so that the stock synchronised in Dealcar can also be served as a product feed for ChatGPT Ads without additional work.
You can see how Google VLA works at dealcar.io/dealcar-boost and check the ChatGPT Ads integration at dealcar.io/dealcar-boost/chatgpt-para-concesionarios.
Frequently asked questions
Can I use the same stock feed for VLA and ChatGPT Ads?
The field format is different: VLA has specific fields for automotive (fuel, mileage, year, transmission) and ChatGPT Ads uses a generic product format. The base inventory is the same, but you need to generate two separate feeds adapted to the format of each platform.
Which has a higher return for a small dealership with a limited budget?
Without proprietary ChatGPT Ads data in Spanish automotive yet available, VLA is the safer bet due to the maturity of the channel and the availability of sector benchmarks. ChatGPT Ads can be tested with a small investment (200-300 euros per month) in parallel to gather proprietary data before scaling.
Can Google VLA and ChatGPT Ads cannibalise each other's leads?
In practice, no. They capture the buyer at different moments of the process and within different contexts. A buyer may have seen both channels in their search process, but that is not cannibalisation: it is coverage at multiple points of the funnel.
Table of Contents
The stage of the buying process captured by each channel
How targeting works in each one
Ad formats and how they appear
The product feed: differences between the two systems
How performance is measured on each channel
Compared costs: what to expect from each
What kind of leads each channel generates
When to use only VLA, when to use only ChatGPT Ads, and when to use both
The question dealerships ask most: do ChatGPT Ads replace portals?
Dealcar and the two channels
Frequently asked questions
The stage of the buying process captured by each channel
The most important difference between the two channels is not technical: it is the moment they reach the buyer.
Google Vehicle Ads intercepts active, specific search. The buyer has already more or less decided what they want: they search for "second-hand diesel 2020 Volkswagen Golf Barcelona" or "used family SUV under 20000 euros". They already know the type of car, sometimes the specific model, and they are comparing options available now. This is high and specific intent.
Read also how the car search process is changing with ChatGPT in 2026.
ChatGPT Ads intercepts the conversational consideration phase. The buyer is still defining what they need: they explain to ChatGPT that they have two children, drive 50 km a day, do not want petrol, and have an £18,000 budget, and they ask for recommendations. This buyer has not decided on the model yet, but they have a very clear need and are receptive to specific options.
Neither of the two stages is better in the abstract. They are different. VLA captures the buyer closer to the closing of the deal. ChatGPT Ads captures them higher up the funnel, when they can still be influenced by the options they see.
How targeting works in each one
Google VLA works by vehicle model: the advertiser uploads the stock feed and Google shows the cars when someone searches for that specific model or related terms. Targeting is by keywords and search signals that Google automatically interprets.
ChatGPT Ads works through context hints: descriptions in natural language of the buyer's situations and needs. There are no keywords. The advertiser writes something like "person looking for a reliable second-hand family car for intensive use" and OpenAI's system determines in which conversations to show that ad based on the context of the active conversation.
Check what ChatGPT Ads context hints are and how to use them to sell cars.
The practical consequence is that VLA has more predictability: the dealership knows their ads appear when someone searches for that model. ChatGPT Ads has more contextuality: the ad appears in conversations where the context is relevant, even if the buyer has not typed the name of the model.
For a dealership with varied stock, VLA allows targeting by model and managing different bids for high-demand cars. ChatGPT Ads requires defining the buyer profile well, rather than the product.
Ad formats and how they appear
In Google VLA, the ad appears at the top of the search results in a horizontal carousel format, featuring the car's photo, make, model, price, and the dealership's name. The user is on the Google results page, viewing multiple cars side by side.
In ChatGPT Ads, the ad appears below the response generated by the model, in a separate block labelled as sponsored. The user is reading the AI's response to their question, and the ad appears right after. The format includes a vehicle image, title, description, and link to the detail page.
The difference in context matters. On Google, the user expects to see ads interspersed with results: it is part of the familiar format. On ChatGPT, the user is in the middle of a conversation, and the ad arrives just as they have received a relevant response for their decision. This can lead to greater receptivity or more interruption, depending on how the ad is designed.
The product feed: differences between the two systems
Both channels work with a vehicle product feed, but the format and fields are different.
Google VLA has a specific feed format for automotive with sector-specific fields: number of passengers, fuel type, body style, transmission, year of manufacture, mileage, and condition (new or used). Google validates that the feed meets the requirements and processes it to display the cars in the VLA carousel.
ChatGPT Ads uses OpenAI's generic product feed: item_id, title, description, url, image_url, price, availability, seller_name, and seller_url. There are no specific fields for vehicles in the current official format. The title and description are the fields the AI uses to understand what car it is and when to display it.
Read also how to prepare the car product feed for ChatGPT Ads.
This means that the quality of the texts in the feed matters more in ChatGPT Ads than in VLA. In VLA, the structured format of specific fields tells Google exactly what kind of car it is. In ChatGPT Ads, the title and description are the information the system uses to determine contextual relevance.
A good description for a car in the ChatGPT Ads feed is not "Volkswagen Golf 2020 diesel 115HP": it is "Volkswagen Golf 2020 automatic diesel, 67,000 km, serviced at official workshop, ideal for daily city use or long journeys, C-label, transfer included".
How performance is measured on each channel
Google VLA has a mature measurement system integrated with Google Ads and Google Analytics. Conversions are set up in Google Ads, tracking works with the Google pixel, and performance data by vehicle model is available with good granularity.
ChatGPT Ads has its own measurement system in OpenAI Ads Manager: impressions, clicks, CTR, CPC, and conversions with the OpenAI pixel. The channel is new, and historical data for benchmarking is scarce. For traffic reaching the website from ChatGPT Ads without the pixel installed, the oppref=chatgpt parameter in the URL allows it to be identified in Google Analytics.
Attribution is more complex in ChatGPT Ads because the buying cycle can be long: the buyer sees the ad on ChatGPT, visits the website days later from another channel, and buys over the phone. In VLA, the cycle can also be long, but tracking is more mature.
Read also how to measure ChatGPT Ads leads and sales in a dealership.
To compare the actual performance of the two channels, the most useful metric is cost per qualified lead, not CPC. A lead coming from ChatGPT Ads that reaches the detail page of a specific car with contact details can be worth more than a cheap VLA click that bounces in 10 seconds.
Compared costs: what to expect from each
Google VLA does not have a standard public cost because it works by auction. In Spain, the typical CPC for used car dealerships varies by model and area, but is in a range of between 0.30 and 1.50 euros per click on used car searches.
ChatGPT Ads has a recommended bid of 3 to 5 dollars per click according to OpenAI, with a CPM starting from approximately 25 dollars. This cost per click is higher than VLA's, but the audience receiving the ad is in a context of greater receptivity. If the click-to-lead conversion rate is higher in ChatGPT Ads, the cost per lead can be comparable or lower.
In the first experiments published by advertisers in other sectors, the CTR of ChatGPT Ads is between 1.5% and 3.2%, compared to a 0.35% average for conventional display. Specific automotive data for Spain is not available yet as the channel has just opened.
What kind of leads each channel generates
The lead coming from Google VLA already knows what model they are looking for and is comparing available options. It is a lead with high and specific intent: if they reach the detail page of a specific car, they are very likely to be considering that car.
The lead coming from ChatGPT Ads might be in an earlier decision stage. They have described their need to the AI, received a response with recommendations, and clicked on an ad for a car that fits what the AI suggested. It is a contextually qualified lead but with a potentially longer decision cycle.
In practice, this translates into different sales processes. The VLA lead might be ready to visit the car this week. The ChatGPT Ads lead might need more information and more follow-up before committing to a visit.
When to use only VLA, when to use only ChatGPT Ads, and when to use both
For a dealership not using either of the two channels, the recommended order is to start with VLA. The intent is more direct, tracking is more mature, and there is more reference of what works in the sector. Once VLA is optimised and generating leads with a known return, add ChatGPT Ads as a second layer.
Also check the complete Google VLA guide for dealerships.
For a dealership already using VLA with good results and wanting to diversify acquisition channels, ChatGPT Ads is the natural next channel to test. The initial test budget can be 200 to 400 euros per month to gather enough data in 30 days.
For a dealership that has a very specific stock profile (cars of a single segment, very specific geographic area, customer with a defined profile), ChatGPT Ads can be especially efficient because the context hint can describe that buyer very accurately and the feed only contains relevant cars.
Both channels together make sense when the budget allows and when there is capacity to manage leads from two channels with different follow-up processes. It makes no sense to activate ChatGPT Ads if VLA leads are already not managed well due to lack of capacity.
The question dealerships ask most: do ChatGPT Ads replace portals?
No, at least not in the short term. Portals (Coches.net, AutoScout24, Wallapop) remain the channel where most used car buyers in Spain start their search. They have decades of indexed content, millions of vehicle detail pages, and a consolidated audience.
ChatGPT Ads does not compete with portals on the same playing field: it is not a marketplace where the buyer filters by model and price. It is a conversational channel where the ad appears when someone is asking for help to decide. They are different parts of the search process.
What might change in the medium term is the proportion of buyers starting on ChatGPT instead of Google or a portal. If that percentage grows, the dealership that already has a presence on ChatGPT Ads will have an advantage over the one that has not yet tested the channel.
Dealcar and the two channels
Dealcar automatically manages the stock feed for Google VLA through Dealcar Boost: when a car enters the stock, it appears in VLA ads; when it is sold, it disappears. Dealcar is developing the same integration for ChatGPT Ads, so that the stock synchronised in Dealcar can also be served as a product feed for ChatGPT Ads without additional work.
You can see how Google VLA works at dealcar.io/dealcar-boost and check the ChatGPT Ads integration at dealcar.io/dealcar-boost/chatgpt-para-concesionarios.
Frequently asked questions
Can I use the same stock feed for VLA and ChatGPT Ads?
The field format is different: VLA has specific fields for automotive (fuel, mileage, year, transmission) and ChatGPT Ads uses a generic product format. The base inventory is the same, but you need to generate two separate feeds adapted to the format of each platform.
Which has a higher return for a small dealership with a limited budget?
Without proprietary ChatGPT Ads data in Spanish automotive yet available, VLA is the safer bet due to the maturity of the channel and the availability of sector benchmarks. ChatGPT Ads can be tested with a small investment (200-300 euros per month) in parallel to gather proprietary data before scaling.
Can Google VLA and ChatGPT Ads cannibalise each other's leads?
In practice, no. They capture the buyer at different moments of the process and within different contexts. A buyer may have seen both channels in their search process, but that is not cannibalisation: it is coverage at multiple points of the funnel.
