• Meta Ads

Dynamic remarketing on Facebook and Instagram

Dynamic Facebook remarketing is about showing a person the same position they just viewed on your site, not another "come back to us" banner. It all depends on one condition: the ID of the product in the browsing event must match the ID in the Meta catalog. They run away quietly - there are numbers in the cabinet report, but someone else's goods are in the banner. Therefore, we start not with campaigns, but with a match check: two or three days, free of charge.

See how it works
Price
after a free audit
Guarantee
30 days after project sign-off
Setting period
5-14 business days after the match is confirmed
after a free audit
Price
30
Guarantee

days after project sign-off

5-14
Setting period

business days after the match is confirmed

free event and directory match check, 2-3 business days
Where do we start?
lingerie store: the remarketing module gives the product_id separately for Google and Facebook
Measurement 01.08.2026
self-defense store: 68 addresses in the directory, from the Meta Pixel ad stack only
Measured on 07/31/2026
the frequency of impressions per user and the share of impressions for those who have already bought
What we measure after launch
from your advertising account, we will not resell it
Show budget

Free event and directory matching check

Dynamic ads substitute exactly the product that a person was looking at in the banner. For this to work, the id in the view event must exactly match the id in the directory. The most frequent breakdown is right here, and it looks like "the ad shows the wrong thing".

What we measure

  • Coincidence of product identifiersWe compare what the view event transmits and what is in the directory. In one of the stores we measured, the remarketing module gave the product_id for several advertising systems at once - in this way, the match is achievable.
  • Complete chain of eventsViewing the product, adding it to the cart, starting the checkout, purchasing. If the middle is missing, the segments will be rough.
  • Directory statusAvailability, price, link, image. An ad shows an item that is out of stock just as readily as one that is available.
  • Exclusion of buyersDoes the segment of those who have already bought be subtracted. The most common defect of the channel is to spend months showing a person a product that he has already received.
  • Segment retention periodsOne term for all is a typical mistake. Different windows are needed for consumables and for furniture.
  • Frequency of impressionsLet's see how many times a week one person sees your ads. This is where remarketing turns into an annoyance.

What you get

  • ID match report: what's being passed, what's in the directory, and where the discrepancy is.
  • A list of events that are missing from the chain.
  • A scheme of segments and exclusions for your purchase cycle.
  • Conversation for 30 minutes on the document.

Timeline: 2-3 working days

Why is it free

Because a match check takes two to three days, and a dynamic campaign on discrepant IDs shows random products and spends the budget silently - everything looks fine in the report.

What's next

If there is a match, we give an estimate for setting up campaigns. If not, separately to correct the transmission of events or the directory.

Short form: your contact and site URL

  • Contract, act and 30-day warranty

    Every project gets a written contract: scope, deadlines, amount, acceptance procedure. After delivery — act and invoice, then 30 calendar days of warranty.

  • Sole proprietor & bank transfer

    The contractor is a registered sole proprietor. Payment by invoice with closing documents.

  • Rights & access — yours

    Code, design and materials transfer to you after full payment. Domain, hosting, repository and analytics are registered to you.

  • Client portal instead of email chains

    During the project you get access to a portal: contracts, invoices, acts and project status in one place.

  • European clients

    Among our work — projects for Norway, Bulgaria, Moldova and Spain.

  • Verifiable numbers

    Every case in the portfolio comes with a link to a live site and a technical measurement.

  • Audit first, then pricing

    There is no price list on the site intentionally: the scope of the same work differs multiples between clients.

  • We say "no" when unsure

    If the task isn't ours or the deadline is unrealistic — we tell you upfront.

What affects the price

Why two seemingly identical tasks are priced differently

  • Identifier matchThe main factor. If the viewing event and the catalog already speak the same code, the work is reduced to campaigns and segments. If not, first the transmission of the event on the site or the field in the catalog is corrected, and this is another volume.
  • The state of the chain of eventsWhen all four steps are present—view, cart, checkout, purchase—segments are built at once. When there is only a view and a purchase, the middle has to be added, and without it you cannot collect the abandoned basket.
  • How many product groupsOne retention period for the entire catalog is done in one hour. Different windows for consumables, equipment, and furniture are separate sets, separate creatives, and separate validation.
  • Directory sourceA ready-made feed that is actually updated is one story. The export, once collected by hand and since then untouched, must first be brought to a suitable condition.
  • Browser events or server events tooThe pixel is set quickly. The Conversions API adds event key matching, deduplication, and checking that the server event doesn't duplicate the browser event.
  • Who makes creative decisionsWhen there is a person who approves texts and layouts, the rotation follows a schedule. When there is not, the ads burn out, the frequency increases, and the channel begins to be scolded for what it was not given.
Who it's for

Situations where this service delivers results

Scenario 1 of 4

The catalog is large, and the person leaves after looking at a few cards

The more items, the less chance that the buyer will remember exactly where he saw what he needed. In the underwear store we measured, each size and color is listed on a separate card — 115,456 product items, and "similar product" does not work here: a person needs a 70B, not a 70A. The dynamic format substitutes exactly that position in the banner. An honest limit: he will just as conscientiously replace it if the site already redirects to another.

We'll review your situation in a free audit
What's included

Complete list of work and what you get as a result

  • We check the ID of the product in the viewing event with the ID in the catalog and match them - after that, the banner shows the position that the person was looking at, and not the adjacent one.
  • We connect the catalog to dynamic campaigns and collect sets of products according to the rules — displays are controlled by the availability and price filter, not a manual list.
  • We divide the audience according to the depth of the action: viewed, put in the cart, started checkout. Each segment receives its own message, not one for all.
  • We build the abandoned cart segment as the difference between "placed, but not bought" — those who have already paid do not appear in the impressions.
  • We exclude buyers and those who have issued a return — the budget stops going to people who no longer need the product.
  • We set different retention periods for different product groups — consumables and furniture no longer live behind the same window.
  • We set up cross-selling: to those who have already bought, we show related items, and not the same product a second time.
  • Limiting the frequency of impressions per user is exactly the setting that separates reminders from harassment.
  • We duplicate key events through the Conversions API with deduplication by common event ID — some browser events do not reach Meta.
  • We prepare a monthly report with a breakdown by segments and product groups, where conversions after a click and after an impression are separate.
When this service isn't right

What's not included — so there are no surprises at delivery

  • Advertising budget - pay it directly from your account.
  • Assembling a product feed from scratch is a separate job with a separate deadline.
  • Promises of percentage returns from abandoned carts.
  • Newsletters by mail and in messengers.
  • Bypassing restrictions related to the user's opt-out of tracking.
Process steps

Transparent stages with approval at every step

Total duration:5–14 days

  1. Event and directory matching check

    2-3 working days, free of charge

    We compare what the viewing event transmits and what is in the catalog, we check the completeness of the chain and the state of availability. The output is a list of works, based on which the estimate is calculated.

  2. Fixed event transmission

    1-3 working days

    We bring the identifier to the same format, add the steps that are missing, and raise the Conversions API with deduplication by the common identifier of the event.

  3. Catalog and sets of goods

    1 working day

    We connect the catalog, collect sets according to the rules, turn off from the displays items that are not available and those that you do not want to return: one-time, seasonal, withdrawn from sale.

  4. Segments, exclusions, retention periods

    1-2 working days

    Abandoned cart as a difference, separate windows for product groups, exclusion of buyers and returns. Here we set the limit of the frequency of impressions per user.

  5. Creatives and launch

    1-3 working days

    We prepare templates for dynamic formats, divide messages according to the time period after the visit, start up and in the first days we look at the structure of impressions, and not at conversions - the sample is still too small.

  6. The first cycle of reporting and editing

    1-3 working days

    After two or three weeks, we look at the frequency, the share of impressions for those who have already bought, the cost of the returned purchase separately from the initial purchase. We fix the retention windows and ad rotation.

Did not find your case?

Describe how it works on your side — we will tell you whether “Dynamic remarketing in Meta” fits and what it means in your situation. No brief and no call: one question, one answer.

Cases

Tasks and results in numbers — all metrics measured by us

online retail underwear store, catalog with hundreds of thousands of items

Task
Return the buyer to exactly the position he was looking at. Each size and color has a separate card, so "same product" and "similar product" are different things here: a person needs a 70B, not a 70A.
Solution
Store on OpenCart 3 with a remarketing module that gives the product_id separately for Google and Facebook; e-commerce event transmission is enabled in GA4, with GTM and Google Ads running nearby. We took the technical indicators of the site and compared the volume of the catalog using two independent methods.
Result
Measure 08/01/2026: 115,456 product items by search and 112,646 unique product addresses in the site map, a discrepancy of 2.4%. The identifier for the two advertising systems is transferred, that is, a match is achievable. The main risk found is for the dynamic channel: in a sample of 150 random product addresses, 50 (33%) return 301, and all 50 lead to a different position — a different size or color. On the scale of the catalog, this is about 37,000 addresses, on which a click from the banner leads to the wrong product.

a specialized online store of self-defense equipment

Task
Understand whether a dynamic return channel makes sense with a small catalog.
Solution
Before talking about the campaigns, the state of measurement and the volume of the catalog were taken: which counters are there, how many addresses are in the site map, how quickly the page is delivered.
Result
Measure 07/31/2026: the site map has 48 products and 18 categories — 68 addresses in total, the smallest catalog of the lot. From the advertising stack, only Meta Pixel stands, no Google analytics are detected at all, zero JSON-LD blocks. First server response 494.7 ms. The conclusion for the channel is straightforward: there is a pixel, but there is almost nothing to return to - 48 items exhaust the segment in a few days, and the system begins to show the same products to the same people.

online store of professional cosmetics, catalog of over 5,000 items

Task
Launch the return of buyers with advertising in Meta where the assortment is constantly updated.
Solution
We measured what the channel starts from: counters on the main page, the state of the site map, the actual volume of the catalog.
Result
Measurement 07/31/2026: 5,561 products (55 pages of 100 items plus 61 for the last one). There is no counter on the main one - no GTM, no gtag, no Facebook Pixel; the sitemap returns an HTTP 200 with an empty body. That is, there is a separate job before setting up dynamic campaigns: without events on the site, there are no segments, and without a working site map and a feed, there is nowhere to get the catalog.
Technologies & integrations

What we build on and what it connects to

Stack

  • Meta Ads Manager - Campaigns, product sets and impression rate cap live here
  • Meta Pixel — events on the site; some of them are cut by blocker or consent mode, so it is incomplete by itself
  • Conversions API — server copy of events; closes what didn't fly from the browser, but needs a common event ID, otherwise the statistics will double
  • The Meta Product Catalog is a source of pricing, availability, links and images; is updated exactly as often as it is unloaded
  • The Events Manager is where you can see if the event and directory IDs match
  • Dynamic ads and product sets — item selection rule instead of a manual list
  • Stack Limit: Everything is kept on the directory. There is no catalog - there will be no dynamic formats, and no setting in the cabinet will replace this

Integrations

  • catalog of Meta products
  • Meta Pixel
  • Conversions API
  • GA4 - e-commerce
  • CRM - reconciliation of actual orders
Dynamic ads versus regular retargeting with a single banner

How this option differs from the alternative

What does a person see?the item she was looking at with its price and availability from the catalog
Launch conditionneed a directory and match IDs in events
Work on creativesthere is one template, the products are placed by themselves - there is no need to draw a banner for each position
Products out of stockare disabled by a set-level rule if the directory is updated in time
When something else is smarterwhen the catalog is live and there are many items
What we need from you

We can't start without this — best to prepare in advance

  1. Access to advertising cabinet, business account and directory.
  2. Working product catalog or consent to its connection.
  3. Access to the site, if you have to edit the transmission of events.
  4. The average decision-making cycle by product group.
  5. List of goods that should not be returned at shows: one-off, seasonal, discontinued.
  6. Monthly budget for the return channel.

If something is missing — let us know, we'll help you gather it or do it as a separate task.

FAQ

Most frequently asked questions — with concrete answers

Why is the ad not showing the products I was looking at?

Almost always due to a discrepancy in identifiers. The viewing event transmits one code, another is in the directory — and the system substitutes what it finds, or even random items from the set. This cannot be configured in the office: it is necessary to edit either the transmission of the event on the site or the field in the catalog. There is also a second reason, already after the click: in the underwear store we measured, 50 out of 150 verified product addresses (33%) gave a 301 for a different size or color - the banner showed the correct position, but the landing led to the wrong place. The check takes two to three days and distinguishes these cases clearly.

Can I show ads to cart abandoners?

Yes, and this is the channel's strongest scenario. But an important detail: the segment should be built as a difference - "put in the cart, but did not buy." If you simply take everyone who placed in the basket, buyers will also be in the shows. We additionally expose a narrower window: a person who dropped the basket yesterday and a person who did it a month ago react differently. We do not mention the percentage of returns: Baymard records about 70% of abandoned baskets on the market, but this is the average for the industry, and some of the baskets are thrown deliberately - neither a letter nor a banner will return them.

How not to turn remarketing into harassment?

Limitation of frequency and rotation of messages. The same ad shown ten times a week is more annoying than helpful. We set a limit of impressions per user, divide creatives by time after the visit and exclude those who have already received an order. The order here is important: first the frequency limit and exclusion of buyers, and only then the extension of the retention window. Conversely, you'll get an increase in impressions without an increase in orders.

Do you need a separate directory for Meta or will a Google feed work?

The data source may be the same, but the requirements are different: the set of mandatory fields and the identifier format are different. We usually build one layer of export and make two feeds from it - so prices and availability remain consistent between sites. It is cheaper than maintaining two independent exports, and saves from the classics, when the price is new on one site, and still old on the other.

The pixel already exists - why else the Conversions API?

The browser event does not always occur: blocker, consent mode, interrupted transition, browser restriction on cookies. Server Copy closes this hole. But only if both events share the same event ID - otherwise you'll get double statistics and an overestimated channel. This is the most common connection error: API is enabled, deduplication is not. Fair boundary: The server can't see what the user has given up, and no scheme can get around it.

Does dynamic remarketing work with low traffic?

Worse than with a large one, but the threshold here is lower than in search lists. The problem of a small store is different: the segments quickly run out, and the system begins to show ads to the same people in a circle. In the self-defense store we measured, the catalog has 48 products and 18 categories, a total of 68 addresses; with such a volume, there is almost nothing to return. In such a situation, we either extend the retention window, or we advise you to first invest in attracting new traffic.

How to calculate the result if a person saw an advertisement, but did not click?

Through separate accounting of conversions after impression and after click. They have different values, and lumping them into a single figure is a way to overvalue a channel. In the report, we show them separately, and next to them we put data from your accounting system: how many orders actually came in during the period. GA4 and the cabinet give a trend, not an accounting truth - the reconciliation with the cash register is up to you, and this is the most honest picture available.

What to do with goods that are not available?

Exclude them from sets. A dynamic campaign without an availability filter will faithfully show a person something that cannot be bought - and you will pay for a click that ends in disappointment. We set a rule at the level of a set of products and separately check that the catalog is updated often enough for this rule to work: the filter for outdated balances does not save from anything.

Give the website address and access to the cabinet - we'll start with a match check, not an estimate

In 2-3 working days, we will return with a document: what the view event conveys, what lies in the catalog, where the discrepancy is, what steps are missing in the chain, and what the segment diagram should look like for your purchase cycle. Then there is the cost and term by stages.

From measured casesMeasure 07/31/2026: the site map has 48 products and 18 categories

View cases
  • Reply within 2 hours
  • No commitment
  • We work under a contract

There is no price list on the site on purpose: the same work differs several times over between two clients, and a “from” figure explains nothing in that case. First a free audit — we count your pages, duplicates and speed — then we name the sum and the deadline and fix both in the contract.