Table of Contents
What is a context hint and how does it differ from a keyword
How OpenAI uses context hints in matching
The 280-character limit and how to make the most of it
Most common mistakes when writing context hints
The formula for writing an effective hint
12 examples of context hints for used car dealerships
How to organise ad groups by hint for a diverse stock
Context hints for budget-friendly cars
Context hints for SUVs and family cars
Context hints for premium and late-model used cars
Context hints for buyers migrating from new to used
How to measure which hints perform best
Dealcar and ChatGPT Ads
Frequently asked questions
What is a context hint and how does it differ from a keyword
A keyword is a string of text that the advertiser wants to target: when someone types "second-hand diesel Golf", the system shows the ad if the dealership bids on that keyword. The matching is literal: word for word.
A context hint is a semantic description of the conversations where the ad might be relevant. It is not what the buyer types: it is the situation the buyer describes. OpenAI's system evaluates the meaning and intent of the active conversation and compares it with the hint to decide if the ad is relevant.
The practical difference is profound. A keyword captures a specific search. A context hint can capture dozens of different conversations where the same buyer describes the same need in different ways: "I'm looking for a car to take my kids to school", "I need a family car that fits in the work car park", "I want to swap my saloon for something bigger for weekend trips". Those three phrases do not share a single keyword, but they describe the same buyer profile and the same decision point.
Read also ChatGPT Ads vs Google Vehicle Ads for dealerships.
OpenAI is explicit in its documentation: context hints guide matching but are not exact match rules and do not guarantee delivery in specific conversations. The system evaluates the hint along with the ad headline, description, landing page, and the context of the conversation. A highly precise hint in an ad with a poorly relevant landing page can perform worse than a broader hint with a well-built vehicle detail page.
How OpenAI uses context hints in matching
The ChatGPT Ads system works on a relevance-weighted auction. When there is an active conversation where an ad can appear, the system evaluates all candidate ads and scores them based on two dimensions: the bid amount and the contextual relevance.
Contextual relevance is calculated by cross-referencing the ad group's context hint with the content of the active conversation, along with the ad headline and description, and the content of the landing page. A more relevant ad can win the position over another with a higher bid, which means writing a precise hint not only improves impression quality: it can also reduce cost per click by winning on relevance instead of bid.
See the complete guide to ChatGPT Ads for dealerships
This mechanism has an important consequence: the context hint does not work alone. The hint, the ad headline, the description, and the landing page are a unit. If the hint describes a buyer looking for a spacious family car and the landing page leads to the listing of a compact city car, the consistency between elements is broken and the system penalises it in relevance.
The 280-character limit and how to make the most of it
OpenAI limits the context hint to 280 characters per ad group. It is enough space to precisely describe the buyer profile and the conversation context, but not to write everything that could be relevant.
One way to optimise this space is to focus on three elements: who the buyer is (demographic or usage profile), what need they have (what they use the car for, what problem they want to solve), and what stage of the decision they are at (comparing options, looking for alternatives to buying new, ready to purchase). With these three elements in 280 characters, you describe a concrete buying situation that the system can recognise in conversations.
What not to include in the hint: the dealership's name, the exact price of the cars, financing conditions, or any information that belongs in the ad and not in the buyer's profile. The hint describes the buyer, not the offer.
Most common mistakes when writing context hints
The most common mistake is writing a hint as if it were a keyword: "used cars", "second-hand SUV London", "buy used Golf". These phrases do not describe a conversation or a buyer profile: they are search terms that OpenAI's system does not process as effective hints because they lack the intent context that the matching needs.
The second mistake is being too generic: "people who want to buy a car". This description applies to millions of conversations with very different intents (someone buying a new Tesla, someone looking for a £5,000 runabout, someone who hasn't yet decided whether to buy or keep their current car) and generates low-relevance impressions.
The third mistake is including the dealership's name or the city as the main hint. Geographic targeting is configured at the campaign level, not in the hint. Writing "Manchester dealership" or "Birmingham car sales" in the hint mixes two different types of targeting and reduces the contextual precision of the hint.
The fourth mistake is using a single hint for the entire stock. A dealership with 30 cars of very different profiles (budget city cars, family SUVs, premium cars) that uses the same hint for all of them is being imprecise: the hint that works for the buyer of an £8,000 Seat Ibiza is not the same as the one for a £28,000 BMW 3 Series.
The formula for writing an effective hint
A useful formula to structure an automotive context hint combines four elements: who the buyer is, what life situation they are going through, what they need from the car, and what stage of the decision they are in.
The structure is: "[Buyer profile] who [specific situation or need] and [is looking for / is comparing / is about to decide on] [type of car or main feature]."
Applied to a used car dealership: "Family with children that needs more space and is comparing used SUVs or family cars between £15,000 and £22,000 with a good maintenance history." In 185 characters, this hint describes who the buyer is, what they need, what type of car they are looking for, and in what price range.
Not all hints need all four elements. For highly specific car profiles (a convertible, a large 4x4, an electric car of a specific model), the type of car already filters enough and the hint can focus on the decision point and the usage situation.
12 examples of context hints for used car dealerships
The following hints are organised by buyer profile and are editable according to the dealership's specific stock.
For the first-time car buyer: "Young person looking for their first car on a tight budget, prioritising reliability and low running costs over performance or extras."
For the growing family buyer: "Family that has just had a second child and needs to swap their compact saloon for a car with a larger boot and space for child seats."
For the professional with intensive use: "Sole trader or commercial traveller who drives more than 20,000 miles a year and is looking for a reliable, fuel-efficient used diesel to reduce cost per mile."
For the urban buyer facing emission zones: "Person living in a city with a Clean Air Zone who needs a car with a compliant Euro 6 petrol or ULEZ-exempt engine to drive without charges."
For those migrating from new to used: "Person who wanted to buy a new car but, after seeing prices, is considering a nearly-new car under two years old as a more affordable alternative."
For the family SUV buyer: "Family that uses the car for daily school runs and weekend trips and wants a spacious, automatic used SUV that is ULEZ compliant."
For the premium second-hand buyer: "Buyer with a budget between £25,000 and £40,000 who prefers a 2 or 3-year-old premium used car over a brand-new mid-range car."
For the used electric vehicle buyer: "Person interested in switching to an EV but who finds the price of new ones too high and is evaluating used electric cars with good real-world range."
Read also the demand for used EVs and the Auto+ Plan in 2026.
For the small fleet business: "Small business that needs to add one or two cars to its fleet and is looking for ex-lease used vehicles with documented service history."
For the buyer replacing their current car: "Owner of an older diesel car who is thinking of changing it before emission restrictions further limit their city driving."
For the budget city car buyer: "Person looking for a used city car under £10,000 for daily urban use, prioritising easy parking and low fuel consumption."
For the buyer coming from a classifieds portal: "Person who has spent weeks browsing cars on online portals, has saved several favourites, and is about to decide which one to view in person first."
How to organise ad groups by hint for a diverse stock
For a dealership with mixed stock, the most efficient structure is to create different ad groups for clearly distinct buyer profiles, filtering the feed by ads_metadata so that each ad group only serves cars relevant to that profile.
Read also how to prepare your car product feed for ChatGPT Ads.
A structure of three ad groups covers most medium-sized dealership stocks. The first for budget range (cars under £12,000, city cars and compacts) with hints aimed at first-time buyers, buyers on a tight budget, and urban usage. The second for mid-range (£12,000 to £25,000, SUVs, family cars and saloons) with hints aimed at families, high-mileage professionals, and buyers migrating from new to used. The third for premium range (over £25,000, luxury cars and late-model nearly-new) with hints aimed at premium used buyers and businesses.
This segmentation allows you to manage different budgets and bids for each segment, analyse which buyer profile converts best, and optimise the hints of each ad group independently without the results of one segment skewing the others.
Context hints for budget-friendly cars
Buyers of cars under £10,000 or £12,000 have very specific needs that can be precisely described in the hint: limited budget, reliability over performance, easy to park, low maintenance costs. A hint for this segment that works well is specific about the budget and the usage profile.
What to avoid in budget range hints is mentioning specific brands (the system may interpret it as too narrow a hint) and terms like "bargain" or "lowest price" (which describe the offer, not the buyer). The hint describes who the buyer is, not what promotion the dealership is running.
Context hints for SUVs and family cars
SUVs and family cars have a very defined buyer profile: families with children, people who need a large boot, buyers who prioritise space and comfort over sporty driving. Hints for this segment can be highly descriptive of the buyer's life situation.
An effective tactic for SUVs is to combine the body style with the reason for use ("trips with kids", "dogs", "long commutes") and the expected vehicle condition ("with service history", "ex-lease"). These elements make the hint more specific and help the system find conversations where the buyer is describing exactly that situation.
Context hints for premium and late-model used cars
The buyer of premium used cars has a different decision-making logic: they are not looking for the cheapest car but the best balance between quality, brand, and price. Hints for this segment should reflect that logic: a buyer who values driving experience, safety and comfort extras, and who prefers a premium brand over a new mid-range car.
A common mistake in high-end hints is using price as the sole differentiator. A hint that says "people looking for cars over £30,000" describes the budget but not the motivation. A hint that says "buyer who prefers a 3-year-old BMW or Audi with a warranty over a brand-new mid-range car for the same price" describes the motivation and decision point with much higher precision.
Context hints for buyers migrating from new to used
This segment is particularly valuable for a used car dealership: the buyer who wanted a new car but reconsidered after seeing the prices. This profile's conversations in ChatGPT are highly recognisable: they compare the cost of buying new with what they get in a used model, ask about warranties for used cars, and seek reassurance about the vehicle's condition.
See also how the used car buyer decides in the UK in 2026.
A hint for this profile: "Person who had the budget for a brand-new mid-range car but is comparing if they are better off with a 1 or 2-year-old nearly-new model from a premium brand with more spec."
How to measure which hints perform best
OpenAI Ads Manager shows metrics by ad group, allowing you to compare the performance of each hint in terms of impressions, CTR, and conversions. For a dealership with three ad groups using different hints, direct comparison reveals which buyer profile responds best to the ads.
What cannot be directly measured is what portion of the CTR is due to the hint versus the ad headline, image, or landing page. These elements work as a unit, and the ad group's performance reflects the combination of all of them. To isolate the effect of the hint, you must keep the headline, image, and landing page constant and only change the hint between variants of the same ad group.
The most important signal for evaluating a hint is not CTR but the quality of the leads it generates. A hint that attracts buyers who match the profile of the dealership's stock generates leads that convert. A hint that is too broad may generate many clicks from buyers with different intents who do not end up purchasing.
See also how to measure ChatGPT Ads leads and sales at a dealership.
Dealcar and ChatGPT Ads
Dealcar is developing stock integration with ChatGPT Ads, including the automatic generation of suggested context hints based on the profile of the dealership's inventory. If you want to be among the first to activate the channel when the integration becomes available, see dealcar.io/dealcar-boost/chatgpt-para-concesionarios.
Frequently asked questions
How many context hints can be included in an ad group?
OpenAI allows one context hint per ad group, with a 280-character limit. To cover multiple buyer profiles, you must create different ad groups with different hints.
Does the context hint determine exactly which conversations the ad appears in?
No. OpenAI is explicit that the hint guides matching but does not guarantee delivery in specific conversations. The system evaluates multiple signals, including the hint, ad headline, description, landing page, and conversation context. A highly precise hint in an ad with high internal consistency (aligned hint, ad, and landing page) performs better than a precise hint with a generic landing page.
Do I need to change the hint frequently to optimise?
It depends on results. If the ad group has a consistently low CTR for two or three weeks with sufficient impression volume, testing a variation of the hint with different emphasis (more focused on the decision point, usage profile, or vehicle type) can improve performance. Do not change the hint before having enough data to compare.
Can I include the name of the dealership or city in the context hint?
The dealership name in the hint rarely adds contextual value: it describes the offer, not the buyer. The city or geographic area is better configured at the campaign targeting level, not in the hint. A hint that says "buyers in London looking for a family SUV" mixes geographic and contextual targeting inefficiently.
Table of Contents
What is a context hint and how does it differ from a keyword
How OpenAI uses context hints in matching
The 280-character limit and how to make the most of it
Most common mistakes when writing context hints
The formula for writing an effective hint
12 examples of context hints for used car dealerships
How to organise ad groups by hint for a diverse stock
Context hints for budget-friendly cars
Context hints for SUVs and family cars
Context hints for premium and late-model used cars
Context hints for buyers migrating from new to used
How to measure which hints perform best
Dealcar and ChatGPT Ads
Frequently asked questions
What is a context hint and how does it differ from a keyword
A keyword is a string of text that the advertiser wants to target: when someone types "second-hand diesel Golf", the system shows the ad if the dealership bids on that keyword. The matching is literal: word for word.
A context hint is a semantic description of the conversations where the ad might be relevant. It is not what the buyer types: it is the situation the buyer describes. OpenAI's system evaluates the meaning and intent of the active conversation and compares it with the hint to decide if the ad is relevant.
The practical difference is profound. A keyword captures a specific search. A context hint can capture dozens of different conversations where the same buyer describes the same need in different ways: "I'm looking for a car to take my kids to school", "I need a family car that fits in the work car park", "I want to swap my saloon for something bigger for weekend trips". Those three phrases do not share a single keyword, but they describe the same buyer profile and the same decision point.
Read also ChatGPT Ads vs Google Vehicle Ads for dealerships.
OpenAI is explicit in its documentation: context hints guide matching but are not exact match rules and do not guarantee delivery in specific conversations. The system evaluates the hint along with the ad headline, description, landing page, and the context of the conversation. A highly precise hint in an ad with a poorly relevant landing page can perform worse than a broader hint with a well-built vehicle detail page.
How OpenAI uses context hints in matching
The ChatGPT Ads system works on a relevance-weighted auction. When there is an active conversation where an ad can appear, the system evaluates all candidate ads and scores them based on two dimensions: the bid amount and the contextual relevance.
Contextual relevance is calculated by cross-referencing the ad group's context hint with the content of the active conversation, along with the ad headline and description, and the content of the landing page. A more relevant ad can win the position over another with a higher bid, which means writing a precise hint not only improves impression quality: it can also reduce cost per click by winning on relevance instead of bid.
See the complete guide to ChatGPT Ads for dealerships
This mechanism has an important consequence: the context hint does not work alone. The hint, the ad headline, the description, and the landing page are a unit. If the hint describes a buyer looking for a spacious family car and the landing page leads to the listing of a compact city car, the consistency between elements is broken and the system penalises it in relevance.
The 280-character limit and how to make the most of it
OpenAI limits the context hint to 280 characters per ad group. It is enough space to precisely describe the buyer profile and the conversation context, but not to write everything that could be relevant.
One way to optimise this space is to focus on three elements: who the buyer is (demographic or usage profile), what need they have (what they use the car for, what problem they want to solve), and what stage of the decision they are at (comparing options, looking for alternatives to buying new, ready to purchase). With these three elements in 280 characters, you describe a concrete buying situation that the system can recognise in conversations.
What not to include in the hint: the dealership's name, the exact price of the cars, financing conditions, or any information that belongs in the ad and not in the buyer's profile. The hint describes the buyer, not the offer.
Most common mistakes when writing context hints
The most common mistake is writing a hint as if it were a keyword: "used cars", "second-hand SUV London", "buy used Golf". These phrases do not describe a conversation or a buyer profile: they are search terms that OpenAI's system does not process as effective hints because they lack the intent context that the matching needs.
The second mistake is being too generic: "people who want to buy a car". This description applies to millions of conversations with very different intents (someone buying a new Tesla, someone looking for a £5,000 runabout, someone who hasn't yet decided whether to buy or keep their current car) and generates low-relevance impressions.
The third mistake is including the dealership's name or the city as the main hint. Geographic targeting is configured at the campaign level, not in the hint. Writing "Manchester dealership" or "Birmingham car sales" in the hint mixes two different types of targeting and reduces the contextual precision of the hint.
The fourth mistake is using a single hint for the entire stock. A dealership with 30 cars of very different profiles (budget city cars, family SUVs, premium cars) that uses the same hint for all of them is being imprecise: the hint that works for the buyer of an £8,000 Seat Ibiza is not the same as the one for a £28,000 BMW 3 Series.
The formula for writing an effective hint
A useful formula to structure an automotive context hint combines four elements: who the buyer is, what life situation they are going through, what they need from the car, and what stage of the decision they are in.
The structure is: "[Buyer profile] who [specific situation or need] and [is looking for / is comparing / is about to decide on] [type of car or main feature]."
Applied to a used car dealership: "Family with children that needs more space and is comparing used SUVs or family cars between £15,000 and £22,000 with a good maintenance history." In 185 characters, this hint describes who the buyer is, what they need, what type of car they are looking for, and in what price range.
Not all hints need all four elements. For highly specific car profiles (a convertible, a large 4x4, an electric car of a specific model), the type of car already filters enough and the hint can focus on the decision point and the usage situation.
12 examples of context hints for used car dealerships
The following hints are organised by buyer profile and are editable according to the dealership's specific stock.
For the first-time car buyer: "Young person looking for their first car on a tight budget, prioritising reliability and low running costs over performance or extras."
For the growing family buyer: "Family that has just had a second child and needs to swap their compact saloon for a car with a larger boot and space for child seats."
For the professional with intensive use: "Sole trader or commercial traveller who drives more than 20,000 miles a year and is looking for a reliable, fuel-efficient used diesel to reduce cost per mile."
For the urban buyer facing emission zones: "Person living in a city with a Clean Air Zone who needs a car with a compliant Euro 6 petrol or ULEZ-exempt engine to drive without charges."
For those migrating from new to used: "Person who wanted to buy a new car but, after seeing prices, is considering a nearly-new car under two years old as a more affordable alternative."
For the family SUV buyer: "Family that uses the car for daily school runs and weekend trips and wants a spacious, automatic used SUV that is ULEZ compliant."
For the premium second-hand buyer: "Buyer with a budget between £25,000 and £40,000 who prefers a 2 or 3-year-old premium used car over a brand-new mid-range car."
For the used electric vehicle buyer: "Person interested in switching to an EV but who finds the price of new ones too high and is evaluating used electric cars with good real-world range."
Read also the demand for used EVs and the Auto+ Plan in 2026.
For the small fleet business: "Small business that needs to add one or two cars to its fleet and is looking for ex-lease used vehicles with documented service history."
For the buyer replacing their current car: "Owner of an older diesel car who is thinking of changing it before emission restrictions further limit their city driving."
For the budget city car buyer: "Person looking for a used city car under £10,000 for daily urban use, prioritising easy parking and low fuel consumption."
For the buyer coming from a classifieds portal: "Person who has spent weeks browsing cars on online portals, has saved several favourites, and is about to decide which one to view in person first."
How to organise ad groups by hint for a diverse stock
For a dealership with mixed stock, the most efficient structure is to create different ad groups for clearly distinct buyer profiles, filtering the feed by ads_metadata so that each ad group only serves cars relevant to that profile.
Read also how to prepare your car product feed for ChatGPT Ads.
A structure of three ad groups covers most medium-sized dealership stocks. The first for budget range (cars under £12,000, city cars and compacts) with hints aimed at first-time buyers, buyers on a tight budget, and urban usage. The second for mid-range (£12,000 to £25,000, SUVs, family cars and saloons) with hints aimed at families, high-mileage professionals, and buyers migrating from new to used. The third for premium range (over £25,000, luxury cars and late-model nearly-new) with hints aimed at premium used buyers and businesses.
This segmentation allows you to manage different budgets and bids for each segment, analyse which buyer profile converts best, and optimise the hints of each ad group independently without the results of one segment skewing the others.
Context hints for budget-friendly cars
Buyers of cars under £10,000 or £12,000 have very specific needs that can be precisely described in the hint: limited budget, reliability over performance, easy to park, low maintenance costs. A hint for this segment that works well is specific about the budget and the usage profile.
What to avoid in budget range hints is mentioning specific brands (the system may interpret it as too narrow a hint) and terms like "bargain" or "lowest price" (which describe the offer, not the buyer). The hint describes who the buyer is, not what promotion the dealership is running.
Context hints for SUVs and family cars
SUVs and family cars have a very defined buyer profile: families with children, people who need a large boot, buyers who prioritise space and comfort over sporty driving. Hints for this segment can be highly descriptive of the buyer's life situation.
An effective tactic for SUVs is to combine the body style with the reason for use ("trips with kids", "dogs", "long commutes") and the expected vehicle condition ("with service history", "ex-lease"). These elements make the hint more specific and help the system find conversations where the buyer is describing exactly that situation.
Context hints for premium and late-model used cars
The buyer of premium used cars has a different decision-making logic: they are not looking for the cheapest car but the best balance between quality, brand, and price. Hints for this segment should reflect that logic: a buyer who values driving experience, safety and comfort extras, and who prefers a premium brand over a new mid-range car.
A common mistake in high-end hints is using price as the sole differentiator. A hint that says "people looking for cars over £30,000" describes the budget but not the motivation. A hint that says "buyer who prefers a 3-year-old BMW or Audi with a warranty over a brand-new mid-range car for the same price" describes the motivation and decision point with much higher precision.
Context hints for buyers migrating from new to used
This segment is particularly valuable for a used car dealership: the buyer who wanted a new car but reconsidered after seeing the prices. This profile's conversations in ChatGPT are highly recognisable: they compare the cost of buying new with what they get in a used model, ask about warranties for used cars, and seek reassurance about the vehicle's condition.
See also how the used car buyer decides in the UK in 2026.
A hint for this profile: "Person who had the budget for a brand-new mid-range car but is comparing if they are better off with a 1 or 2-year-old nearly-new model from a premium brand with more spec."
How to measure which hints perform best
OpenAI Ads Manager shows metrics by ad group, allowing you to compare the performance of each hint in terms of impressions, CTR, and conversions. For a dealership with three ad groups using different hints, direct comparison reveals which buyer profile responds best to the ads.
What cannot be directly measured is what portion of the CTR is due to the hint versus the ad headline, image, or landing page. These elements work as a unit, and the ad group's performance reflects the combination of all of them. To isolate the effect of the hint, you must keep the headline, image, and landing page constant and only change the hint between variants of the same ad group.
The most important signal for evaluating a hint is not CTR but the quality of the leads it generates. A hint that attracts buyers who match the profile of the dealership's stock generates leads that convert. A hint that is too broad may generate many clicks from buyers with different intents who do not end up purchasing.
See also how to measure ChatGPT Ads leads and sales at a dealership.
Dealcar and ChatGPT Ads
Dealcar is developing stock integration with ChatGPT Ads, including the automatic generation of suggested context hints based on the profile of the dealership's inventory. If you want to be among the first to activate the channel when the integration becomes available, see dealcar.io/dealcar-boost/chatgpt-para-concesionarios.
Frequently asked questions
How many context hints can be included in an ad group?
OpenAI allows one context hint per ad group, with a 280-character limit. To cover multiple buyer profiles, you must create different ad groups with different hints.
Does the context hint determine exactly which conversations the ad appears in?
No. OpenAI is explicit that the hint guides matching but does not guarantee delivery in specific conversations. The system evaluates multiple signals, including the hint, ad headline, description, landing page, and conversation context. A highly precise hint in an ad with high internal consistency (aligned hint, ad, and landing page) performs better than a precise hint with a generic landing page.
Do I need to change the hint frequently to optimise?
It depends on results. If the ad group has a consistently low CTR for two or three weeks with sufficient impression volume, testing a variation of the hint with different emphasis (more focused on the decision point, usage profile, or vehicle type) can improve performance. Do not change the hint before having enough data to compare.
Can I include the name of the dealership or city in the context hint?
The dealership name in the hint rarely adds contextual value: it describes the offer, not the buyer. The city or geographic area is better configured at the campaign targeting level, not in the hint. A hint that says "buyers in London looking for a family SUV" mixes geographic and contextual targeting inefficiently.
