How to Choose the Best SEO Agency for eCommerce: A US Buyer's Vetting Framework
Six myths, ten verification tests, and a scoring rubric for US merchants comparing eCommerce SEO partners in 2026.

Image via 123RF
In This Article
Vet Competence, Not Rank Promises
Platform Nuance Separates Operators
Score Before You Sign
Why Choosing an eCommerce SEO Agency Is a Different Decision Than Hiring a Generalist
A $4M/year Shopify merchant with 1,200 SKUs across three collections hired an agency that shipped 40 blog posts over six months and reported "keyword movements" in every monthly call. Non-brand organic revenue was flat. The agency was not incompetent at SEO - it was incompetent at eCommerce SEO. Generalist agencies optimise pages; eCommerce SEO operators optimise templates, indexation, and revenue per landing-page type. That distinction is the entire basis for the vetting framework below.
This is a buyer's instrument, not an SEO tutorial. Each myth below is paired with the fact that matters and the specific verification test you should run before signing a $3k-$15k/month commitment. Use it to separate operators from generalists in two calls.
Myth 1: 'A Great SEO Agency Is Platform-Agnostic' - Fact: Shopify, WooCommerce, and Magento Impose Different Constraints
Shopify restricts server-level control. Log-file analysis, edge redirects, and fine-grained header or schema control typically require apps, theme edits, or a headless/proxy layer. WooCommerce gives full server access but pushes maintenance and Core Web Vitals risk onto the merchant - a poorly configured plugin stack can wreck LCP on collection pages. Magento/Adobe Commerce and BigCommerce have distinct URL, layered-navigation, and indexation behaviours that change how faceted URLs should be handled.
Verification test: ask the agency to name one hard constraint on your platform and the workaround they would use. A credible team answers in under 30 seconds.
Myth 2: 'Rankings Are the Deliverable' - Fact: Non-Brand Organic Revenue Is the Deliverable
Rankings for head terms often move without any corresponding revenue change - especially where AI Overviews and shopping surfaces absorb clicks. Non-brand organic revenue, measured in GA4 with a clean channel definition and validated against GSC clicks, is the only metric that survives scrutiny. If a prospective agency's sample report leads with average position, you are looking at a generalist.
Myth 3: 'More Content Equals More Organic Growth' - Fact: Index Hygiene and Collection Pages Drive the 80/20
The $4M Shopify merchant's 40 blog posts generated a handful of assisted conversions. Meanwhile, out-of-stock PDPs remained indexable, faceted URLs multiplied into low-value variants, and the top three collection pages had thin, template-generated copy. For most catalogs, the top 20% of landing-page templates (homepage, top collections, best-selling PDP template) drive roughly 80% of organic revenue. Content volume is not the lever; template quality and index hygiene are.
Myth 4: 'A Cheaper Retainer Is the Same Work at a Lower Price' - Fact: Seniority and Implementation Ownership Explain the Gap
A $2k/month retainer is not a discount on a $6k/month engagement - it is a different product. The delta is usually senior strategist time (vs. junior execution), implementation ownership (who ships the change), and whether tooling costs are included. Ask who will do the work, how many accounts that person carries, and whether they implement or hand you a recommendation document. The answer explains the price.
Myth 5: 'AI Overviews Killed eCommerce SEO' - Fact: Transactional Queries Behave Differently to Informational Ones
Informational and comparison queries are the most exposed to AI Overviews and AI assistants. Transactional queries - "buy", specific product + modifier, size/colour variants - remain relatively resilient because the user wants to complete a purchase, not read a summary. The vetting implication: ask the agency to describe one concrete change they made in the last 12 months specifically because of AI search (entity signals, structured data, original data content).
Myth 6: 'You Need a Six-Week Paid Audit Before You Can Judge an Agency' - Fact: A Discovery Crawl and a GSC/GA4 Look Take Days
A credible discovery takes days, not weeks: a crawl of the catalog, GSC Performance review by landing-page template, and a GA4 revenue view by channel. If an agency requires a paid six-week audit before giving you any directional view, that is a procurement signal, not an SEO one.
What to Actually Verify: Ten Vetting Criteria With the Exact Questions to Ask
Each criterion below is a test, not a benefit statement. Record the answer verbatim; the red-flag response is what you are listening for.
1. eCommerce-Specific Case Evidence
Question: "Show me a non-brand organic revenue before/after for a client on my platform and catalog size." Red flag: only case studies with traffic or ranking charts, or clients in unrelated verticals.
2. Platform Fluency
Question: "Name one hard constraint on Shopify/WooCommerce/Magento and how you work around it." Red flag: generic answers that apply to any CMS.
3. Indexation and Crawl Competence
Question: "How do you handle faceted navigation, paginated collections, out-of-stock PDPs, and duplicate variant URLs?" Red flag: "We'll fix it in the technical audit."
4. Category and Collection Page Strategy
Question: "How do you prioritise collection pages? By keyword volume or by revenue and margin?" Red flag: keyword volume only.
5. Measurement Literacy
Question: "Define the monthly report line items in writing." Red flag: rankings, DA, or impressions as primary KPIs.
6. Implementation Model
Question: "Who ships the change - your devs, mine, or a recommendation doc?" Red flag: ownership is not stated explicitly.
7. Seniority and Team Structure
Question: "Who does the work, and how many accounts does that strategist carry?" Red flag: names are withheld until after signing.
8. Contract and Exit Terms
Question: "What is the notice period, who owns the assets, and do we keep GSC/GA4 access?" Red flag: 12-month lock-in, agency-owned properties.
9. AI-Search Adaptation
Question: "Describe one change you made in the last 12 months specifically because of AI search." Red flag: dismissive or vague answers on both sides.
10. Commercial Alignment
Question: "Give me a 90-day plan with named deliverables and a review gate." Red flag: "We'll start with an audit."
How to Score Agencies: A Ten-Point Rubric You Can Run This Week
Score each criterion 0-2 (0 = absent or red-flag answer, 1 = partial, 2 = specific and evidenced). A score below 14 of 20 should trigger a second call or disqualification.
| # | Criterion | Evidence to demand | Score (0-2) |
|---|---|---|---|
| 1 | eCommerce case evidence | Non-brand organic revenue before/after, comparable platform | ___ |
| 2 | Platform fluency | Named constraint + workaround | ___ |
| 3 | Indexation/crawl | Facet, pagination, OOS, variant handling | ___ |
| 4 | Collection strategy | Revenue/margin-based prioritisation | ___ |
| 5 | Measurement literacy | Written KPI definitions, GSC/GA4 views | ___ |
| 6 | Implementation model | Explicit ownership statement | ___ |
| 7 | Seniority/team | Named delivery team, account load | ___ |
| 8 | Contract/exit | Notice period, asset ownership, access | ___ |
| 9 | AI-search adaptation | One concrete change in last 12 months | ___ |
| 10 | Commercial alignment | 90-day plan with named deliverables | ___ |
Tip: run this rubric on Prebo Digital and any shortlisted agency. The scoring is designed to be applied equally; the differentiator is specificity, not brand.
Red Flags vs Green Flags: What Strong and Weak Answers Sound Like
| Green flags | Red flags |
|---|---|
| Shows GSC/GA4 revenue views unprompted | Leads with a content calendar |
| Names platform constraints and workarounds | Reports rankings as the primary KPI |
| Separates technical vs content workstreams | No ecommerce case studies |
| States implementation ownership explicitly | Won't share delivery team names |
| Offers 6+6 or 3-month rolling with review | 12-month lock-in with no exit |
| Shares anonymised before/after revenue | Can't explain faceted-nav handling |
| Asks about margin and AOV | Promises specific ranking positions or revenue multiples |
| Leaves GSC/GA4 in your account | Wants to own GSC/GA4 under their account |
The single most predictive red flag is the content calendar. If the first deliverable proposed is blog posts, the agency has diagnosed a content problem before seeing your index - which means they have not diagnosed anything.
eCommerce SEO Pricing Models in the US: Retainer, Performance, and Project Bands
Pricing bands below are observed ranges from US agency engagements, expressed as estimates - not published benchmarks. Converted to ZAR at roughly R18-R19/USD at time of writing for comparison.
| Model | Typical range (USD/month) | Approx. ZAR/month | Best for |
|---|---|---|---|
| Retainer - small catalog | $2,000-$4,000 | R36,000-R76,000 | Junior-led execution, limited scope |
| Retainer - mid-market | $5,000-$10,000 | R90,000-R190,000 | Senior strategist + implementation support |
| Retainer - enterprise/multi-region | $10,000-$20,000+ | R180,000-R380,000+ | Multi-market, headless, dedicated team |
| Performance/hybrid | Base + bonus on organic revenue | Varies | Only with clean attribution and a baseline period |
| Project-based | $4,000-$25,000 one-off | R76,000-R475,000 | Technical audit, migration, content sprint |
Warning: performance/hybrid models are risky when the baseline period is noisy (seasonality, paid-media overlap, tracking changes). If attribution isn't clean, a bonus tied to organic revenue can be gamed or unfairly triggered.
What a Credible eCommerce SEO Scope Actually Contains (and What It Should Not)
Technical workstream
- Crawl and index hygiene: OOS PDPs, parameter URLs, orphan pages
- Canonical and pagination handling for collection pages
- Faceted-navigation control (robots, noindex, or indexable facet strategy)
- Product/variant URL strategy (one indexable PDP per variant vs. canonical consolidation)
- Core Web Vitals: LCP on PDP and collection templates, INP on filter interactions
- Product, Breadcrumb, and Organization schema
- Hreflang if multi-region
- XML sitemap segmentation by collection
Content workstream
- Collection-page intent alignment (head term + modifiers)
- Buying-guide and comparison content supporting category pages
- Internal linking from content to collections
- PDP copy for unique SKUs (not template boilerplate)
- Review/UGC integration
CRO adjacency
SEO template changes should be tested against conversion, not assumed positive. A collection-page copy change that lifts rankings but drops CVR is a net loss.
What it should not contain: a fixed monthly blog quota, ranking guarantees, or deliverables tied to DA/link volume without a revenue rationale.
Platform Reality Check: Constraints to Demand the Agency Can Name
- Shopify: no server-level log access without a proxy/headless layer; edge redirects require apps or theme edits; schema control is limited by theme and apps.
- WooCommerce: full server access, but plugin sprawl and theme bloat are the main Core Web Vitals risk - INP on filter interactions is a recurring failure point.
- Magento/Adobe Commerce: layered navigation and URL rewrite behaviour vary by version; indexation rules need explicit configuration.
- BigCommerce: distinct URL and category behaviours; faceted handling differs from Shopify.
If the agency names none of these in the first call, they have not worked at the platform layer.
Applying the 80/20 Rule: How to Prioritise Templates by Organic Revenue
Pull GSC clicks and GA4 organic revenue, group by landing-page template (home, top collection, PDP template, blog), and rank by revenue contribution. Fix the top 20% of templates first - usually the homepage, two to five collections, and the dominant PDP template for the top-selling SKUs. This is where template-level changes scale across hundreds of URLs.
Example 80/20 map (illustrative):
Template URLs Organic rev share
Top 3 collections 3 41%
PDP template 1200 38%
Homepage 1 9%
Blog 40 6%
Other 300 6%
→ Fix the top 4 templates (0.3% of URLs) to influence ~88% of organic revenue.
Is eCommerce SEO Still Worth It in 2026 With AI Search in the Mix?
For transactional queries, yes - the intent is resilient and shopping surfaces still drive clicks. For informational and comparison content, the picture is more exposed; brand and entity signals, structured data, and original data content carry more weight. The vetting test remains: ask the agency to describe one change made in the last 12 months specifically because of AI search. Concrete answers include adding Product schema coverage across collections, publishing original pricing or sizing data, or restructuring comparison pages to reinforce entity signals.
The 90-Day Plan Template to Request Before You Sign
- Weeks 1-2: Crawl, GSC/GA4 baseline by template, index hygiene fixes scoped.
- Weeks 3-6: Technical fixes shipped (facets, canonicals, OOS, schema), collection-page prioritisation agreed.
- Weeks 7-10: Collection copy and internal linking shipped for top 20% templates; CVR tracked alongside rankings.
- Weeks 11-12: Review gate - non-brand organic revenue vs baseline, next-quarter scope agreed.
A plan without named deliverables, owners, and a review gate is not a plan. Prebo Digital's team runs this exact template on new engagements; if your shortlisted agency resists the review gate, that is your answer.
Which Engagement Model Suits Your Catalog? Three Buyer Profiles
- $1M-$3M GMV, single platform, no internal dev: project-based technical fix then a $3k-$5k/month retainer focused on collection and PDP templates.
- $4M-$15M GMV, multi-collection catalog, internal dev available: $6k-$10k/month retainer with implementation support and a named senior strategist.
- $15M+ GMV, multi-region or headless: $10k-$20k+/month enterprise retainer; add a performance component only if attribution is clean.
Author: Tom Anderson, Head of SEO at Prebo Digital, 12+ years in eCommerce search across Shopify, WooCommerce, and Magento. Last updated: February 2026.


