Google Ads Ad Customizers in 2026: Dynamic Prices, Locations and Countdowns in RSAs
You manage an account with 60 service areas and 400 SKUs. You want the exact price and the city name in the ad copy — that is 24,000 text combinations. Nobody writes those by hand, and generic copy with no numbers loses the click to competitors who show them. Google Ads ad customizers exist precisely for that gap: one responsive search ad that fills itself in differently depending on the query, campaign, device and date.
This guide covers how customizers actually behave in 2026: which data types they support, how to build a business data feed, what the syntax looks like, why substitution silently fails, and what to verify before you push anything live.
Google Ads ad customizers vs. DKI vs. IF functions
A customizer is a variable inside your ad text. Instead of a static line like “Nationwide Delivery”, you write a placeholder that resolves at serve time into “Delivery to Austin in 2 Days”. The values live in a business data feed — a table of attributes mapped to campaigns, ad groups, keywords, locations or devices.
Three mechanisms get confused constantly, and the differences matter:
| Mechanism | What it inserts | Data source | Best used for |
|---|---|---|---|
| DKI (dynamic keyword insertion) | The keyword that triggered the impression | Your keyword list | Tight, hand-curated ad groups only |
| Ad customizers | Any value: price, number, text, date | Business data feed or campaign-level values | Prices, inventory, deadlines, locations |
| IF functions | One of two pre-written strings | A condition: device or audience | Different offer for mobile or returning users |
DKI is the riskiest of the three in 2026. It inserts the matched keyword, not the user’s query, and with broad match your ad can end up carrying phrasing nobody on your team approved. Customizers stay under your control because you author every value in the feed.
Simple rule: if the value changes by product, city, deadline or price, use a customizer. If the whole sentence changes by segment, use an IF function. Reserve DKI for narrow brand or model ad groups where you have reviewed every keyword.
Four attribute types, and why the type is not cosmetic
Every customizer attribute has a type, and the type drives formatting at serve time:
- Text — an arbitrary string: city, category, condition (“no deposit required”). Length is hard-capped by the ad slot.
- Number — a whole number: models in stock, review count, delivery days. Decimals get truncated.
- Price — a currency value. Google formats separators for the user’s locale, so store a clean number plus a currency code, not a pre-formatted string.
- Percent — a percentage. The value is inserted without the symbol, so write the “%” in the ad copy itself.
The classic mistake is uploading prices as Text. They then render inconsistently across locales, skip currency formatting entirely, and are harder to audit. The second is using Number for a value with cents — “199.50” becomes “199”, and your ad now understates the price, which is a disapproval risk rather than a clever discount.
Building the business data feed, step by step
The feed is a spreadsheet you upload into the account. In the 2026 interface:
- Go to Tools → Shared library → Business data.
- Click “+” and choose Ad customizer data.
- Download the template — it already contains the targeting columns.
- Fill in attribute name, type, value, and the targeting columns: Target campaign, Target ad group, Target keyword, Target location, Target device, plus Start date / End date and scheduling.
- Upload and wait for processing. Row-level errors surface directly in the upload report.
The critical concept is targeting hierarchy. The most specific matching row wins. If a campaign-level row and an ad-group-level row both apply, the ad group row is used. If nothing matches at all, the customizer resolves to your default text — or, if you did not write one, the ad does not serve.
Example feed for a multi-location service business
| Attribute | Type | Value | Target campaign | Target ad group |
|---|---|---|---|---|
| City | text | Austin | Search — Austin | |
| Days | number | 2 | Search — Austin | |
| Price | price | USD 149 | Search — Austin | Installation |
| City | text | Denver | Search — Denver |
The headline then reads Installation in {=City} in {=Days} Days and resolves to “Installation in Austin in 2 Days”. If you operate in languages with grammatical cases or gendered nouns, plan for that at the feed level: substitution is literal and does not inflect anything. Create separate attributes per grammatical form rather than trying to reuse one.
Syntax and default text
The full form is {=FeedName.AttributeName:default text}. If you have a single feed with unique attribute names, the short form {=AttributeName:default} works.
Default text is not decoration. Here is exactly what happens without it:
- Value found → the value is inserted.
- No value, default set → default is inserted and the ad serves normally.
- No value, no default → the ad is not eligible to serve.
That last case is the number one reason impressions collapse right after a customizer rollout. The account looks healthy, campaigns are enabled, budget is untouched — but half your ad groups have no matching feed row. Write a default for every customizer, and make it a sentence you would be happy to show, not a placeholder like “product”.
Second rule: the default must fit the slot. If headline plus default exceeds the character limit, that variation is rejected and you get the same silent outage. Always measure the worst case, not the average one.
Countdowns: urgency that is actually true
The countdown customizer inserts remaining time until a date: “Sale ends in 3 days”. The basic syntax is {=COUNTDOWN("2026-09-30 23:59:59","en",3)}, where the last parameter is how many days before the date the countdown starts.
There is also GLOBAL_COUNTDOWN, which counts down in a single fixed time zone rather than the viewer’s local time. Use GLOBAL for a nationwide deadline; use the standard version for “ends today” logic.
Three things break countdowns in practice:
- The date passes. That ad variation stops serving. If every headline carried a countdown, the campaign goes quiet. Always keep at least two static headlines in the RSA.
- The window is too long. “29 days left” creates no urgency and eats characters. A 3–7 day window is a reasonable starting point for testing, not a universal constant.
- There is no real deadline. A sale that “ends” every Friday and never ends inflates CTR while conversion rate drops. Smart Bidding learns from conversions, so low-quality clicks actively degrade the model you depend on.
IF functions: different copy by device and audience
IF functions need no feed and live directly in the ad: {=IF(device=mobile,Book in One Tap):Book Online. Everything after the colon is the fallback for all other cases.
The audience variant looks like {=IF(audience IN (Past Customers),Your 15% Loyalty Rate):New Customer Offer. It only fires for lists attached to the campaign or ad group — at minimum in observation mode. If the list is not attached, the condition never triggers and you will spend an afternoon debugging an ad that was never broken. The mechanics of observation versus targeting are covered in our guide to audience observation in Google Ads.
IF functions cannot be nested, and you cannot combine two conditions in a single expression. If you need a device-by-audience matrix, build it with separate ads or separate ad groups instead of inventing syntax.
How customizers interact with RSAs and asset reporting
An RSA assembles combinations from your headlines and descriptions. A customizer is part of one headline, not a separate entity. Practical consequences:
- Do not make all 15 headlines dynamic. If the feed fails, the system has nothing left to serve. A workable split is 3–5 customized headlines plus static ones.
- Pinning plus customizers shrinks the combination pool fast. Pin the price headline only if the offer depends on it, and never pin more than two positions.
- In asset reporting, a customized headline is scored as one asset regardless of how many values passed through it. So the report will not tell you which city converts best — that analysis belongs at campaign and location level. How to read that report beyond the “Best/Low” labels is covered in our piece on Ad Strength and RSA asset performance.
Rolling out at scale and pre-flight checks
The web UI is fine for piloting on one campaign, but account-wide rollouts belong offline. Google Ads Editor bulk operations let you swap headlines across hundreds of ad groups and, more importantly, surface length warnings before anything uploads.
Pre-launch checklist:
- Every customizer has default text, and that text reads as a finished sentence.
- Slot length is validated against the longest value in the feed, not a sample.
- Every ad group has at least two headlines with no substitution.
- Feed rows cover every campaign and ad group where the attribute is referenced — verify by diffing your ad group list against the targeting column.
- Seasonal rows have Start/End dates and a successor row, so values do not vanish after the end date.
- Feed prices match landing page prices. A mismatch is a direct disapproval trigger; see our guide to Google Ads destination requirements.
Keeping the feed fresh
A manual monthly upload guarantees stale prices sooner or later. Automation options, in increasing order of effort:
| Method | Frequency | Requirement | Main risk |
|---|---|---|---|
| Scheduled fetch from URL or Sheets | 1–4× daily | A stable file URL | A silent format error kills the whole feed |
| Google Ads script | On a schedule | Access to the price source | Needs failure monitoring |
| Google Ads API | Near real time | Engineering time | Maintenance cost |
Whichever you pick, configure a failure alert. The dangerous outcome is not an error message — it is silence: the feed never refreshed, ads keep quoting yesterday’s price, and you find out from a customer.
When customizers are the wrong tool
More often than people assume. Customizers add a failure point, so they earn their place only when the dynamic value genuinely changes the click decision:
- Price, when you are competitive or the range is wide.
- Location, when people search with local intent.
- Availability or lead time, when that is the category bottleneck.
- A deadline, when the deadline is real.
If the inserted value does not change behaviour, you have added complexity for nothing. Your time is better spent on asset breadth — see ad assets and extensions — or on cleaning up the search terms report and negative keywords.
Measuring the impact properly
Customizers almost always lift CTR. CTR does not pay invoices. A defensible test looks like this:
- Run a 50/50 split via Google Ads experiments.
- Decide on CPA or ROAS, not CTR and not Ad Strength.
- Run for at least one full conversion cycle plus a stabilisation period. If the bid strategy changed recently, wait out the Smart Bidding learning period first, or you are measuring learning, not customizers.
- Freeze the feed during the test, otherwise the result is not reproducible.
As a directional benchmark from practice — not a promise — price customizers in competitive categories tend to move CTR by single to low double digit percentages relative to baseline, and the effect scales with how good your price actually is. In categories where price is not an advantage, showing it can reduce conversions because unqualified users self-select out before the click. That is still a useful outcome: you keep the budget.
Policy review: why customized ads get disapproved
Review applies to every possible resolved value, not to the template. Each feed row is effectively a separate ad in policy terms, which produces a distinct set of disapproval causes:
- Ad price does not match the landing page price. The most common one, and it usually comes from feed-to-site drift rather than intent.
- Superlatives in the default text. “Best price”, “#1 provider” require third-party substantiation — and defaults get inserted everywhere, including markets where the claim is indefensible.
- Punctuation and capitalisation inside values. Exclamation marks in headlines, ALL CAPS words, stray symbols — all checked on the assembled string, not on the template.
- Countdowns to a date that has passed. Not technically a disapproval, but the variation stops being eligible, which looks identical from the outside.
- Outcome claims in regulated categories. A customizer is not a loophole: anything you cannot write statically, you cannot insert dynamically either.
Practical habit: before uploading, take 10–15 random feed rows and assemble the final headlines by hand. If any combination reads questionably to you, review will reach the same conclusion — just a week later, with your impressions paused.
A 60-location rollout, step by step
Here is how this plays out for a service business: one service, 60 cities, regional pricing, dispatch time varying by branch load.
- Week 1 — structure. Campaigns split by location, or by location clusters sharing the same economics. This is a prerequisite: feed targeting keys off campaign names.
- Week 1 — feed. Four attributes:
City(text),Price(price),Days(number),Offer(text for regional promotions). Sixty rows per attribute, targeted at campaign level. - Week 2 — pilot. Roll into the five highest-traffic campaigns. Each RSA gets three customized headlines and twelve static ones. Every customizer carries a location-free default such as “Technician Visit in 2 Days”.
- Weeks 2–4 — measurement. A 50/50 experiment in the pilot campaigns. Decision metric is CPA, not CTR.
- Week 5 — rollout. If the experiment holds, push headline replacements to the remaining 55 campaigns via Editor.
- Ongoing — synchronisation. Scheduled daily feed fetch plus an alert on processing failure.
The most common breakage in this scenario is campaign renaming after the feed is built — targeting silently stops matching. Make it a rule that renaming a campaign requires a same-day feed edit. The second weak point is a branch closing while its row stays live, so the ad promises service in a city you no longer cover.
Common failure modes
| Mistake | Symptom | Fix |
|---|---|---|
| No default text | Impressions drop sharply after launch | Add defaults everywhere |
| Price stored as text | Inconsistent formatting, disapprovals | Recreate the attribute as price type |
| Countdown on every headline | Campaign goes silent after the deadline | Keep 2+ static headlines |
| Feed never refreshes | Ad price ≠ site price | Scheduled fetch plus failure alert |
| Audience IF with no list attached | Condition never fires | Attach the list in observation mode |
| Default text overflows the slot | Variation silently rejected | Validate against the longest string |
Where to go next
Customizers operate at the ad copy layer. The next layers up are where you show and who you show to: see search themes in Performance Max and demographic targeting and audience exclusions.
If you are scaling paid search and need reliable account infrastructure, look at how Google Ads agency accounts work and what our full service list covers.
FAQ
How is a customizer different from dynamic keyword insertion?
A customizer inserts a value you authored in a feed, so you control every word. DKI inserts the matched keyword, which gives you less control and a higher chance of awkward or off-brand phrasing.
How many customizers can one ad contain?
There is no meaningful hard cap, but readability and character limits are. One or two substitutions per headline, and three to five customized headlines per RSA, is a sensible ceiling.
Why is my ad with a customizer not serving?
Three usual causes, in this order: no matching feed row and no default text; headline plus default exceeds the character limit; the feed failed processing due to a format error.
Do ad customizers work in Performance Max?
Not in the same form as in search RSAs. PMax drives product-level dynamics through the Merchant Center feed instead, including custom labels in Merchant Center.
Do I need a separate feed per campaign?
No. One feed can serve the whole account; separation happens through the targeting columns. Multiple feeds make sense when different teams own the data or refresh cadences differ.
How quickly do feed updates appear in live ads?
Typically within hours of successful processing. Build a buffer into promotional timing rather than expecting minute-level precision.
Do customizers improve Quality Score?
There is no direct bonus for using them. Indirectly, more relevant copy usually earns higher CTR, and expected CTR feeds into the score. But CTR gains without conversion gains are a false positive.
Can I use a countdown in descriptions?
Yes, the syntax is identical. Urgency works less well there because descriptions are read less often, so keep the countdown in a headline and echo the message statically in the description.
What happens after a feed row’s end date?
The value disappears and the default text takes over. Keep a dateless fallback row at campaign level so seasonal rows expiring never causes an impression gap.
Can I combine an IF function and a customizer in one headline?
Technically possible but fragile and hard to debug. Split them across headlines or ads and compare in asset reporting instead.
Are customizers worth it for a small account?
If you run five campaigns and one offer, almost certainly not — feed maintenance outweighs the gain. They pay off where combinations multiply: many locations, many SKUs, changing prices.
Does the substitution affect ad approval?
Yes. Every possible resolved value must comply with policy, including the default text. Review the feed as if each row were a separate ad, because effectively it is.
