Ecommerce agency cost in India: pricing models and proposal checklist

An ecommerce agency quote is only comparable when the underlying work is comparable. Advertising optimisation, catalogue support and full marketplace operations can all be described as “account management,” even though they require different access, skills, capacity and evidence.

This guide does not publish an invented market average or a Steps Ecommerce rate card. It gives Indian sellers a reproducible way to compare proposals, expose exclusions and decide whether the operating model fits the account.

Why public price ranges are not a reliable benchmark

Current search results use different definitions of ecommerce management. A RiseUp Ecom guide dated 27 October 2026 describes fixed, commission and hybrid arrangements and labels its ranges as figures gathered from public agency pages rather than an independent benchmark. A Deuxen article dated 28 September 2026 combines listing, orders, returns, claims and reconciliation inside one advertised package. The numbers are not comparable until the work, limits and measurement basis are comparable.

Before using any public figure, record:

• whether it covers one marketplace or several;
• the number of seller accounts, brands, SKUs and variations;
• whether advertising, catalogue, account health, returns and settlement work are included;
• whether creative production, taxes, media spend or software are separate;
• the expected volume of launches, cases and catalogue changes;
• reporting and meeting frequency;
• contract duration, notice and handover duties; and
• the date on which the public offer was checked.

A wide range can describe a wide range of work. It does not establish what a particular seller should pay.

Understand the pricing model before comparing the amount

Fixed retainer

A fixed monthly retainer makes the commercial amount predictable. The proposal still needs to define the capacity reserved: marketplaces, SKUs, campaigns, catalogue changes, cases, reports and meetings. It should also state what happens when the workload moves outside those limits.

Percentage of advertising spend

This structure makes the agency fee move with media spend. For advertising-only work, define which accounts and campaign types are included, whether creative and landing-side review are separate, and who may approve budget changes. Ask how the arrangement prevents spend growth from becoming the only route to a larger fee.

Percentage of sales

The sales basis must be written precisely. Ordered value, shipped value, delivered value and net sales after cancellations and returns are different denominators. The agreement should identify the source report and treatment of taxes, discounts, organic sales, marketplace-funded promotions, refunds and later adjustments.

Project or milestone fee

This can fit a finite deliverable such as onboarding, a catalogue migration, a storefront build or a defined audit. Write the acceptance criteria, seller inputs, revision limits, dependency dates and post-delivery support before work starts.

Performance-linked or blended model

Define the baseline, data source, attribution window, exclusions and payment trigger. Both sides should be able to reproduce the calculation from the same exports. A performance label without a signed denominator is not a measurement method.

No model is automatically suitable or unsuitable. The useful question is whether its incentive, evidence and included work match the seller's objective.

Convert every proposal into the same scope table

Create one comparison sheet and require each provider to mark every line as included, excluded, limited or seller-owned. Avoid ticks without an explanation.

• Marketplace operations: seller accounts, marketplaces, order issues, account health and case ownership.
• Catalogue: new SKUs, variations, bulk files, suppression fixes, attributes and approval flow.
• Advertising: channels, campaign types, optimisation cadence, search-term work and budget authority.
• Creative: images, video, A+ or storefront work, revision rounds and source-file ownership.
• Inventory and pricing: replenishment inputs, stock-risk reporting, repricing and promotion approvals.
• Returns and settlements: RTO analysis, payout reconciliation, claim preparation and case follow-up.
• Reporting: metrics, source exports, cadence, commentary and named decision owner.
• Access and security: user roles, credential handling, approval limits and periodic review.
• Exit: notice, file handover, campaign history, creative files and open-case transfer.

Add a limit beside every included item. “Catalogue management” should not conceal a cap on new listings, unsupported variation work or a separate charge for resolving errors.

Require a quote another person can reproduce

A useful quote should connect the commercial model to a written scope and operating assumption. Ask for:

1. the legal entity and seller accounts covered;
2. the objective and baseline that will be used;
3. included workstreams and explicit exclusions;
4. volume assumptions such as active SKUs, campaigns, launches and cases;
5. seller inputs and approval deadlines;
6. the data sources used for reporting;
7. decision rights for budgets, catalogue edits, promotions and cases;
8. deliverables, review cadence and acceptance criteria;
9. the method for handling work outside scope; and
10. notice, access removal and handover requirements.

If one of these changes, the effect on scope and commercial terms should be documented instead of appearing later as an unexplained extra.

Keep account ownership and decision rights with the seller

The seller should retain control of marketplace, advertising, analytics, domain, website and creative accounts. Give the provider only the role needed for the agreed work, record who approved it and review access periodically.

Google's current manager-account ownership guidance says linking an existing Google Ads client account does not give the manager ownership by default. It also recommends making a manager the owner only when those additional privileges are required. Use the access model visible in the current platform rather than sharing a primary login.

Maintain an access register containing:

• platform and account identifier;
• seller-side owner;
• agency user or manager account;
• role, purpose and approval limits;
• date granted and last reviewed; and
• removal and handover status.

Do not store passwords, recovery codes or one-time codes in the register. Named users and attributable partner access are preferable where the platform supports them.

Treat partner badges as one data point

Amazon Ads' current partner-selection guidance says companies listed in its Partner Directory are vetted members of the Amazon Ads Partner Network. It recommends selecting against the advertiser's needs and objectives and treating partner status as one consideration.

That does not mean one directory entry verifies every marketplace service, result claim or member of the delivery team. Check the official profile, supported services and marketplaces, then compare them with the proposal.

Shopify's current India Partner Directory separates providers by service areas such as store setup, development, marketing, analytics and product photography. A provider listed for one discipline should not automatically be treated as verified for unrelated marketplace operations.

For any badge, certification or partnership claim, save the official profile URL and review date. A logo copied into a proposal is not the same evidence.

Define reporting and proof before work starts

A report should connect activity to the seller's stated objective and name its source. Agree the denominator and owner for each important metric.

For marketplace operations, that may include:

• ordered, shipped, delivered, cancelled and returned units;
• net sales on a consistently defined basis;
• contribution inputs after applicable marketplace and fulfilment charges;
• advertising spend, attributed sales and the agreed efficiency measures;
• RTO and return reasons by relevant segment;
• suppressed listings, account-health issues and open cases;
• settlement exceptions, claims filed and documented outcomes; and
• stock risks and decisions requiring seller approval.

Ask for source exports or drill-down records, not only a presentation. A provider should be able to show what changed, who approved it, why it was done and what evidence would reverse the decision.

Do not accept a screenshot, submitted case or attributed-sale figure as proof of cash received or incremental profit. Those are different stages and need different evidence.

Use the first review period to test the operating system

Do not make the first review depend on a promised sales outcome that cannot be isolated from stock, pricing, demand, platform changes and seller decisions. Instead, agree a short operating baseline and inspect whether the work is controlled and reproducible.

The first review should test whether:

1. access, accounts and data sources are documented;
2. catalogue, advertising, return and settlement baselines can be rebuilt;
3. each priority has an owner, due date and expected business effect;
4. material changes have before-and-after evidence;
5. seller approval is recorded for budgets, pricing and policy-sensitive decisions;
6. reports reconcile to source exports;
7. unresolved cases and risks remain visible; and
8. the next actions follow from observed evidence rather than a generic checklist.

This is a process test. It does not prove long-term performance or guarantee a commercial result.

Watch for proposal red flags

Pause the comparison when a proposal:

• guarantees sales, ranking, reinstatement or a fixed advertising outcome;
• quotes a commercial amount without defining the accounts and work covered;
• describes “full management” but leaves catalogue, returns, settlements or case ownership undefined;
• requests a primary password when supported named access is available;
• reports only gross revenue without cancellations, returns, charges or advertising context;
• uses badges that cannot be verified on an official profile;
• treats submitted claims as confirmed recoveries;
• gives the agency control of seller-owned domains, creative files or analytics without a handover rule; or
• has no method for transferring open work and removing access at exit.

A red flag is a prompt for clarification and evidence. Record the answer in the final scope rather than relying on a sales-call assurance.

Final ecommerce agency proposal checklist

Before signing, confirm that the proposal answers these questions in writing:

1. What business problem is the engagement designed to improve?
2. Which accounts, marketplaces, SKUs and workstreams are included?
3. Which tasks remain with the seller or another provider?
4. Who performs the work and who approves material changes?
5. Which source reports define sales, ads, returns and settlements?
6. How are commercial terms calculated, and what can change them?
7. Who owns the accounts, audiences, campaign history, catalogue files and creative sources?
8. What evidence will be used to assess the first review period?
9. What are the notice, access-removal and handover requirements?
10. Which claims, badges and platform relationships can be verified officially?

The useful comparison is not simply “Which agency is cheapest?” It is “Which documented scope gives the account the required work, seller control and reproducible evidence at commercial terms the business can carry?”

Primary platform sources and competing public offers were rechecked on 30 October 2026. The official Google, Amazon Ads and Shopify pages did not display publication dates; their current access and directory descriptions should be rechecked when a seller evaluates a proposal.