A Google Shopping agency is worth hiring when your ad spend is large enough that a 7-15% fee costs less than the margin you are losing, and your catalog is big or volatile enough that feed work is a standing job.
A Google Shopping agency is worth hiring when your ad spend is large enough that a 7-15% fee costs less than the margin you are losing, and your catalog is big or volatile enough that feed work is a standing job. Below roughly $10,000 a month with a stable catalog, the fee costs more than it recovers.
Hire one when two conditions hold at the same time: spend high enough to absorb the fee, and a catalog large or volatile enough that Merchant Center is a real job. General PPC hiring advice tracks one variable, spend. Shopping has two, and the second is where the money leaks.
ZATO Marketing, the most credentialed Shopping specialist on this results page, publishes that its full service targets brands spending $25,000 to $1,000,000 per month, with a lighter tier below $25,000 (ZATO Marketing, 2026). That is one agency's self-segmentation, not an industry standard. It is still the most useful number on the SERP, because it tells you who the sales pages were written about. If you spend $8,000 a month, that is not you.
Shopping sits inside the wider Google Ads optimization playbook, but it carries one structural problem no other Google channel has. That problem decides the hiring question, so start there.
Google Shopping runs on two separate systems, and most hiring mistakes come from treating them as one. We call this the Two-System Split. It turns "should I hire an agency" into a question with an answer.
System 1 is Google Merchant Center, the product data system. Merchant Center enforces the product data specification: seven attributes are required on essentially every product, and a missing one means the product cannot serve (Google Merchant Center, 2026). It also enforces the Shopping ads policies, escalating from disapproved ads to "suspending accounts for repeat or egregious violations" (Google Merchant Center, 2026). Separate login, separate policies, separate ways to go dark overnight.
System 2 is Google Ads, the campaign system: Standard Shopping and Performance Max, asset groups, brand exclusions, bidding, budget, product-level ROAS. Google states the dependency plainly: Shopping ads "use your existing Merchant Center product data (not keywords) to decide how and where to show your ads" (Google Ads, 2026).
So the real question is not agency or no agency. It is who owns System 1 and who owns System 2. The common failure is model 2 below: a general PPC agency takes System 2, and System 1 stays yours by default, usually without anyone saying so out loud. Vysta, an agency describing its own industry, puts it bluntly: agencies that "treat the feed as the client's responsibility" are "operating with a significant blind spot" (Vysta, 2026).
Google agrees, in product form. As of August 2026, Google ships Merchant Center for Agencies, a product "available exclusively for agencies" with its own portfolio dashboard and a diagnostics page that prioritizes fixes by "click potential" (Google Merchant Center, 2026). Google built separate tooling for the Merchant Center side because it is a separate job. For tactics, see what actually moves a Shopping feed.
Published Shopping management fees run 7-15% of ad spend, usually with a flat fee on top. OuterBox states that fees "often include a flat fee plus a percentage of spend, commonly in the 7-15% range" (OuterBox, 2026, read 2026-08-20), and hedges it as planning context, not a quote.
is the ad spend floor the SERP's most credentialed Shopping specialist publishes for full service.
We read the pricing language on every page ranking for this keyword on 2026-08-20. Influencer Marketing Hub puts agency fees "between 5% to 15%," and separately advises allocating "10% to 20% of your total ad spend for the agency" (Influencer Marketing Hub, 2026). Hunter Digital and Softtrix, ranking first and third, publish no figure at all. For a buyer building a budget, that silence is a finding.
The interpretation matters more than the range. A percentage-of-spend fee is regressive for small accounts. Twelve percent of $8,000 is $960 of media budget doing management instead of buying clicks. Twelve percent of $40,000 is $4,800, and at that size the fee buys work that scales with the account rather than with your patience.
Those are illustrative figures, not a client result. Run the same arithmetic on your own spend and see which side the money lands on.
The general ladder of retainers, salaries and subscriptions lives elsewhere: see the general agency vs in-house vs software cost comparison. We also break down what PPC management actually costs.
Four ownership models cover almost every real setup, and each one is defined by who holds System 1 and who holds System 2. Read the second column as Merchant Center first, campaigns second.
| Model | Merchant Center / Campaigns | Cost per month | Best fit | Failure mode |
|---|---|---|---|---|
| Full-service Shopping specialist | Agency / Agency | 7-15% of spend plus a flat fee (OuterBox, 2026-08-20), or flat scoped pricing (ZATO, 2026-08-20) | ZATO publicly targets $25,000 to $1,000,000/mo; large or volatile catalogs | At low spend the fee exceeds the margin it recovers |
| General PPC agency | You, by default / Agency | Similar fee, narrower scope; no page on this SERP publishes a separate rate | Shopping is a minority of spend and the feed is clean | Nobody owns the feed (Vysta, 2026-08-20) |
| In-house or owner-run with a feed tool | You / You | ShoppingFeeder $20/$120/$500 and GoDataFeed $5 per 1,000 SKUs, both per month (vendor pricing pages, 2026-08-20) | Owner-operators with real time and a smaller catalog | Works until the catalog or the owner's calendar grows |
| Self-serve software on campaigns, you keep Merchant Center hygiene | You / Software | Kampaio published tiers $99/$199/$399 per month (2026-08-20) | $3,000-25,000/mo with a manageable catalog | Software does not fix a suspended Merchant Center account or argue a misrepresentation appeal |
Model 1, the full-service Shopping specialist, is the cleanest answer when it fits. One party owns product data and campaigns, so nothing falls between them. You pay for that tidiness.
Model 2, the general PPC agency, is the common trap. Scope looks broad on the proposal and stops at the Google Ads login.
Model 3, in-house with a feed tool, works for owner-operators with genuine time, which is rarer than owners think. DataFeedWatch (which Hunter Digital states it uses), Feedonomics, Channable, GoDataFeed, ShoppingFeeder and Shopify's native Google channel sit in this lane.
Model 4 is self-serve software on campaigns, with Merchant Center hygiene still yours. Kampaio's published tiers are $99, $199 and $399 per month. The limitation in that table cell is real, and we will not bury it: no software argues a policy appeal for you.
Feed operations are ongoing labor, not a setup task that ends. The workload scales with catalog size multiplied by SKU churn multiplied by channel count. That is why a 200-SKU catalog with quarterly changes and a 12,000-SKU catalog with weekly repricing are different businesses on the same Merchant Center.
The recurring jobs are specific: disapproval triage, spec compliance as Google changes requirements, price and availability accuracy, supplemental feeds, seasonal turnover, Merchant Promotions, and appeals. Compliance expands rather than settles. Google now requires disclosure on certain AI-generated or edited assets because "AI regulations in the European Union, India, and New York require" it (Google Merchant Center, 2026). Promotions carry their own cadence: each expires at 183 days and passes editorial review before it displays (Google Merchant Center, 2026).
None of that is difficult work. It is constant, which is a different problem, and it never appears as a line item on a proposal. For the tactical layer, see the full Shopping optimization diagnostic.
Of the 33 public discussions we analyzed on 2026-08-20 about Shopping agency management, Merchant Center appeared as its own named topic in about a fifth. Weak corroboration, but directional.
Suspension risk, not the bidding, is the strongest single argument for paying a specialist. You are buying someone whose full-time job is preventing a zero-revenue week, and whose mistakes do not burn one of your one-to-three appeals.
A good Shopping agency at the right spend level is a straightforwardly good deal, and the reason this results page is full of them is that the model works. Kampaio sells software, so our incentive points the other way, and the answer still does not move: past a certain spend and catalog size, hire the specialist. The argument here is about fit, not about agencies being bad at their jobs.
Two bad-fit signals deserve emphasis. If you cannot name the condition that would get an agency fired, no retainer will produce one for you. And a structural account problem gets inherited, not fixed. A monthly fee does not repair broken tracking, it just puts a nicer dashboard on it.
Before signing, read how to vet a PPC agency. The other path is running Google Ads without an agency.
The difference is not skill, it is scope, and the scope difference is Merchant Center. Shopping ads run on product data, not keywords, so a generalist optimizing bids is working on half the machine.
Six scope differences:
Be precise about Performance Max, because the standard critique has aged. As of August 2026, Google's documentation lists asset-group reporting, channel performance and placement reports as native PMax features (Google Ads, 2026), so the 2023 "black box" complaint is outdated. What has not changed: no PMax surface reports at the individual SKU level, and a retailer's problem almost always lives at the SKU.
Five questions a generalist cannot bluff on a sales call:
Ask them cold. The pauses tell you more than the answers.
Expect 7-15% of ad spend plus a flat fee, per OuterBox's published range, corroborated at 5% to 15% by Influencer Marketing Hub (Influencer Marketing Hub, 2026). The dependency comes second: fees track spend and scope, so the same percentage buys very different work at $8,000 and at $40,000.
Roughly $25,000 a month for full service, the band ZATO Marketing publishes for its own top tier, with a lighter tier below that. At the other end of the market, SmartSites publishes a $1,000+ minimum campaign size. Between $10,000 and $25,000, SKU count and churn decide it.
Whoever you name in the contract. As of August 2026 Google ships Merchant Center for Agencies, and it works by linkage: "agencies must be linked to at least one active Merchant Center account" (Google Merchant Center, 2026). Linkage is not transfer of ownership, so put the responsibility in writing.
Partly. Software runs and monitors the campaign side, System 2. Software does not argue a misrepresentation appeal, and appeals are typically limited to one to three attempts with a seven-day first cool-down. If your binding risk is policy, not bidding, buy the human.
Usually yes, and quickly. Misrepresentation is the most-cited policy behind suspensions, and a suspension persists until a successful appeal (Search Engine Land, 2026). Attempts are limited, so a specialist who has won appeals is worth more than a month of bid tuning.
Yes, if you set it up that way, and you should insist. Both accounts stay registered to your company, with the agency or software granted access rather than title. Kampaio works this way by default. An agency that owns your accounts owns your conversion history.
Google Shopping is two systems, and whoever owns the product data system owns your downside. That is the decision. The retainer is only the invoice attached to it.
Monday action: open Merchant Center, open the diagnostics tab, count the disapprovals and out-of-stock products, and ask who is accountable for that number. If the answer is nobody, you have found the actual problem. Then pick the model that gives that number an owner, whether or not that model has our name on it.
Kampaio runs the campaign side continuously and shows every change, and both accounts stay in your name. Kampaio pricing starts at $99/month.
Vox watches product-level ROAS and moves budget off the SKUs that are not paying, while Merchant Center stays in your name.
See what Kampaio runsResults may vary. This article is informational and does not constitute professional advice. Fee ranges and vendor prices are cited to their original sources and were verified on August 20, 2026; your own spend, catalog and margins will differ, so use the models as a starting frame, not a quote.