Journal · CMP guides

Choosing a cookie consent platform when your agency manages multiple client sites

13 August 20267 min readBy Walter Masseroni Gonzalez, Leon Aves, Daragh McCarthy and Alana Rossi

The right platform is the one where a policy change reaches every site at once and each consent log stays scoped to a single client. Judge candidates on that, not the feature list: who the controller is on each log, whether one configuration replicates or dozens drift, and what a data request costs across different builders.

The feature list is not the decision

Almost every published answer to this question is a ranked list of products, scored on price per domain, white labelling, sub accounts and certifications. Those are real criteria, and the part of the decision a vendor can fit in a table.

What none of them price is the work that lands on your team the day after you buy. A portfolio is not one site multiplied. It is estates on different builders, in different markets, owned by different clients, each auditable on its own. The tool either absorbs that or hands it back as recurring manual work. Four questions settle it.

Some of it is the entry ticket: Consent Mode signals so Google’s tags know what each visitor in the EEA agreed to, a scanner that finds what each site sets, and rules that change by region. Every shortlisted tool should do all three, so none of them decides it. For the ranked list, see our CMP comparison for 2026.

This one carries legal weight and is the most likely to be skipped in procurement. The usual arrangement is that your client controls their visitors’ data and the platform acts as a processor, though not every platform is set up that way, so check its terms. Where agencies get caught is assuming they are a processor too. If you decide why the tracking runs and which purposes the banner declares, you may be a controller yourself, either jointly with your client under Article 26 of the GDPR or separately. The EDPB’s guidelines on controllers and processors say these roles follow what each party actually does, not the label in a contract. So work out the roles from the facts, record them in your contract with each client, then check the platform can express what you agreed.

Operationally that means a consent log attributable to one site and exportable on its own. Handing that record over when a client leaves is a contractual and product question rather than a right they can exercise, which is why it is worth confirming before you sign. And if a platform pools every site's records with no per site boundary, evidencing one client's audit means entering a store holding everyone else's too.

The difference shows when the policy changes.

One policy, replicated

One edit. Every site, same day.

A parent configuration the sites inherit.

  • Site 1
  • Site 2
  • Site 24

Dozens of configurations

Twenty four logins, one forgotten.

Each edited by hand, on its own day.

  • Site 1 on v4
  • Site 2 on v2
  • Site 3 on v4
  • Site 24 missed
The same portfolio, run two ways. The cost on the right is not the licence fee, it is the site you did not get to.

How does a policy change reach every site?

Consent configuration is not something you set once. A client adds a chat widget. A vendor turns out to set an advertising cookie you had categorised as functional. Each of those changes what your banner declares and what it blocks, and each has to reach every site it applies to.

On a single site tool, that is one login per estate, done by hand on a day when other work is also happening. It holds at a handful of sites. Somewhere past that it becomes the thing that silently does not happen, and the drift is invisible because every site still looks fine when you open it. So ask how a change propagates. Is there a parent configuration the sites inherit, or only a copy function that clones once and then lets them diverge? Can you see, in one view, which sites are on the current policy? That last question separates a portfolio tool from a single site tool wearing a multi domain price.

What changes when the estates sit on five different builders?

A real agency roster is rarely one platform. It is some WordPress, a couple of Shopify stores, a Webflow site, something on Squarespace, and one hand built application a developer left behind two years ago. The banner is the easy half of that: most tools render on anything.

The tag stack is where it diverges. Each builder has its own idea of where scripts belong, its own native integrations injecting their own tags, and its own checkout or form flow firing outside the pages you instrumented. A tool that behaves the same across all of them beats one that is excellent on the platform most of your clients do not use. Every extra integration path is another place where behaviour has to be verified separately. That is a different question from white labelling, covered in what actually matters in white label cookie consent.

What does an audit or a data request actually cost?

On a large roster this is the cost that decides the tool, and nobody quotes it. A client's visitor asks what was collected about them, or their legal team asks you to evidence consent on a given date. With a per site boundary and a searchable log, that is a query. On a pooled arrangement it is an exercise: find the account, work out which log belongs to which domain, export it, check you have not included anyone else, repeat across that client's other sites. A few times a quarter across a roster and you have found the real price of the cheaper licence.

What should you check before you commit?

  1. Ask for a single site consent log export

    Not a description of one. Watch someone produce it. If it takes a support ticket, every future audit will too.

  2. Change one category and watch where it lands

    Confirm it reaches a live site without a republish. If the only mechanism is a clone, you are buying drift.

  3. Install it on your least favourite client platform first

    The estate that exposes a tool's limits is the odd one on the roster, not the one you know best.

  4. Check who the log names as controller

    Per site, in writing, matching the roles you agreed with that client. This matters on the day a client leaves and takes their estate with them.

When is a per site tool genuinely the better buy?

Plainly: if you run a handful of sites, all on one builder, all serving one market, a per site tool is simpler and usually cheaper, and everything above buys you nothing.

The crossover is not a site count, it is a rate of change. When the number of sites multiplied by how often your policy changes exceeds what one person can reliably hold in their head, a tooling preference has become an operational risk. Velo is built for that second case: one dashboard across the portfolio, a consent log scoped to each client, and one snippet and one set of category rules carrying across builders. The caveat is that a platform's own native integrations still need checking wherever they inject tags. The model is on the product page.

Earn 25% as a Velo partner

Free to join. No partner fee and no minimum roster.


Agencies and freelancers keep a flat 25% of every plan their clients pay, for as long as they stay.
%
Velo partnersAgencies and freelancers

Free for one site and 1,500 monthly sessions. Pro from €7 a month, plus VAT, for unlimited sites.

Step by step installs for WordPress, Shopify, Webflow, Wix and more.

Common questions

What is the best cookie consent platform for an agency managing multiple client sites?

No single product answer survives contact with a real roster, because what decides it is operational rather than functional. Judge candidates on what one policy change costs across the whole portfolio. Does a parent configuration replicate to every site, or only clone once and then drift? Does each consent log keep a clean per site boundary and record the controller you agreed with that client? Do one snippet and one set of category rules carry across the builders you support, and what does a single data request cost to answer? Consent Mode, scanning and regional rules are table stakes.

Who is the data controller on a client's consent log?

Usually the client, with a caveat agencies often miss. The client controls their visitors' data and the platform usually acts as a processor. But if you decide why the tracking runs and which purposes the banner declares, you may be a controller too, jointly with the client or separately. The roles follow what each party actually does, not the label in a contract. Work them out from the facts and record them in writing. Then check the platform can express it: a log attributable to one site, exportable on its own, and handed over at the end of the relationship because your contract says so.

How do you keep cookie banner settings consistent across dozens of client sites?

With a parent configuration the sites inherit, rather than a copy function that clones settings once. A clone is consistent on the day you make it and diverges from then on, and the drift is hard to see because every site still looks correct on its own. Check whether one view tells you which sites are on the current policy.

Is a per site consent tool ever the better choice for an agency?

Yes. On a handful of sites, one builder, one market, a per site tool is simpler and usually cheaper, and central management is overhead until there is enough to manage. The crossover is a rate of change rather than a site count. Once sites multiplied by how often your policy changes exceeds what one person can hold in their head, how the tool replicates a change becomes the deciding factor.

Keep reading


Free for one site

Banner, scan and consent records for one site, free. Pro from €7 a month, plus VAT, for unlimited sites.