7 Questions to Ask Before Replacing Your Shopify Agency

A theme name or missing app signature can prompt a useful question. It cannot tell you whether a store has an agency, what the agency was hired to do, or whether its work is effective.

Two March 9, 2026 StoreInspect studies provide concrete assessment candidates: a 240,429-record redesign cohort and a separately extracted 240,121-record ICP cohort. Their counts describe stored technology and contact fields, not agency performance.

Some links in this article are affiliate links. We may earn a commission if you purchase through them, at no extra cost to you.

What the evidence covers

The redesign study used stored theme, app-count, traffic-tier, Plus-flag and contact-count fields. The ICP category queries used the latest available app snapshot, without a successful-scan requirement or maximum snapshot age.

Both studies used estimated traffic tiers. Many prospect recipes used the literal condition of a stored traffic tier other than <10k or <50K. That legacy exclusion can admit other labels and should not be described as a verified numeric 50K floor. A positive contact count means a stored contact exists; it does not establish a verified buyer, deliverability or lawful outreach channel.

Public scans can miss native, backend and custom implementations. They also cannot observe contracts, paid acquisition budgets, revenue, conversion rates or an agency's contribution to them. Keep that distinction when using the ICP framework.

1. Does the theme prevent agreed work?

The redesign study matched 57,594 records to a predefined legacy theme-name list. The retained rows included:

That is about 24% of this historical cohort, not all Shopify stores. A theme name does not establish its version, customizations or present capabilities.

The study found 808 legacy-name records passing the traffic exclusion rule. A narrower nine-name query—Debut, Brooklyn, Supply, Minimal, Venture, Simple, Boundless, Narrative and Pop—returned 609 with a positive contact count. Its name list differs from the broader legacy count.

Shopify's migration assessment, checked October 10, 2026, allows a vintage theme to remain when it meets the merchant's needs. Modern theme architecture can make sections and app blocks easier to manage, but migration requires checking compatibility and preserving custom work.

Ask the agency to demonstrate the capability that is blocked, compare an update with migration, and estimate the work. Test mobile navigation, product templates and app compatibility. A concrete maintenance constraint supports a project; a legacy name alone does not support replacing an agency. See the redesign guide.

2. Is there an actual email-program gap?

In the separate ICP latest-snapshot query, 28.5% of 35,493 records in the 50K–200K tier had no email-category match. The under-50K query had its own 202,850-record denominator and a 68.4% unmatched rate. Neither rate covers all higher tiers.

A directly queried email recipe returned 8,142 candidates with no category match, the legacy traffic exclusion and a positive contact count. It did not measure campaign activity, consent, ROI or agency responsibility.

Check signup behavior, consent wording and what happens after a valid test subscription. Ask which platform sends lifecycle messages and which flows are active. Shopify-native or backend-only delivery may leave no recognizable public signature.

If email work is in the contract, review execution and measurement against its agreed goals. If it is outside scope, discuss whether it should be added. Use the email app comparison to evaluate a tool after establishing the task.

3. Does a free theme meet the store's requirements?

The redesign cohort retained this distribution:

The four free-theme counts sum to 3,193. A separate recipe found 3,087 free-theme records with a Plus flag and positive contact count, without a traffic condition. Do not compare it as a subset of the same higher-traffic pool.

Free themes can support successful stores. Paid themes add particular design and merchandising options, but purchase price does not establish faster pages, higher conversion or competent management.

List what shoppers and staff need: useful product information, accessible interactions, suitable collection layouts and maintainable content. Test the current implementation. Compare configuration, targeted development and a replacement theme before choosing a redesign.

4. Can the team demonstrate measurement works?

A public scan cannot prove that advertising is active, inactive or properly measured. A missing Meta signature can reflect consent state, server-side events, a custom implementation or detector coverage. A visible pixel likewise does not prove valid purchase events or current spend.

Ask for a measurement walkthrough using authorized first-party access: consent behavior, one test conversion, browser/server deduplication where relevant, reporting definitions and reconciliation to orders. Check actual campaign status in the advertising account.

Evaluate whether tracking work was agreed, implemented and maintained. Do not infer agency neglect from pixel presence or absence. The pixel-detection guide explains what a public signal can establish.

5. Are important workflows unsupported?

The ICP store-field table averaged 1.5 detected apps under 50K, 3.1 at 50K–200K and 4.8 at 200K–1M. These are cross-sectional averages. They do not establish how many apps a store should have or a growth sequence.

A recipe that used fields stored on each store record returned 11,211 records with zero to two detected apps, the traffic exclusion and a positive contact count. Another analyst-labeled “underinvested” group contained 581 higher-traffic records with zero to three apps. No unmet needs, budget or sales outcomes were measured.

Ask staff which tasks create manual work or errors. Map that workflow to native features, existing systems and any proposed app. Fewer apps may reflect effective native tools or bespoke development; more apps can increase cost and maintenance.

Use the technology-by-traffic guide as context, then assess one concrete bottleneck.

6. Are customers missing useful product evidence?

The ICP study found no reviews-category match in 52.9% of the 50K–200K tier. A directly counted recipe returned 19,995 records with the legacy traffic exclusion and no reviews match. It did not require contact data.

Inspect actual product pages. Reviews can exist through native, custom or unrecognized systems. Some stores have few eligible orders, and some product categories need a different way to answer shopper questions.

If review collection is wanted, assess invitation eligibility, authenticity, moderation and how feedback appears. Establish baseline coverage before proposing a reviews app. No installation-time or immediate-ROI promise follows from the detection count.

7. Does Plus solve the requirements that justified it?

A separate ICP query returned 37,055 Plus-flagged and 203,067 unflagged records, totaling 240,122. Their average detected app counts were 3.7 and 1.4. The stored flag is a detection signal, not a verified billing plan, and the score includes technology inputs.

As checked October 10, 2026, Shopify Plus pricing lists USD $2,300/month on a three-year term or $2,500/month on a one-year term. That subscription does not prove an agency budget or require a particular vendor stack.

Ask which checkout, B2B or organizational requirements led to the plan choice, and whether the agreed implementation meets them. A thin public stack can coexist with substantial internal systems. Use the Plus identification guide to understand public evidence.

Decide using scope and outcomes

Before switching agencies:

1. Restate the contract, deliverables, access and responsibilities.
2. Document a specific unresolved problem with first-party evidence.
3. Ask the agency for its diagnosis, corrective work and measurement plan.
4. Compare delivery quality, communication and results with the agreed scope.
5. Plan account ownership, code access, documentation and handover if replacement is warranted.

A detection study cannot set the acceptable conversion lift or tell you when an agency has failed. Judge outcomes over an appropriate period using consistent reporting, while accounting for seasonality, merchandising, acquisition changes and work done by other teams.

For agencies researching prospects

Use these fields to choose a question, then validate the storefront and buyer. Historical candidate pools overlap and cannot be added into a single agency market. The TAM guide explains joint-filter sizing.

A clearly hypothetical outreach example:

“I saw the Debut theme name on your storefront. Theme names do not tell me which customizations you have. Is maintaining product templates a constraint for your team? If it is, I can outline the update and migration options.”

Another:

“I couldn't identify a public email-platform signature. You may use a native or backend setup. Are lifecycle messages already covered, or is there a workflow you want help with?”

Do not promise a migration lift, infer zero email revenue, or claim a detector proves neglect. Confirm a relevant contact and suitable outreach channel. StoreInspect's Free plan has basic search and 10 reveals; advanced filters and CSV export require a paid plan, as checked October 10. A reveal still needs role and suitability checks.

How many stores need a new agency?

These studies do not measure that. They identify historical technology segments for assessment. The answer depends on the merchant's actual needs, existing work and agency agreement.

Is a modern theme or a larger app stack enough?

No. Assess functional requirements, maintainability and measured customer outcomes. A free theme or compact stack can meet them.