How Agencies Price a Conversational Commerce Add-On

What Do Agencies Charge for a Conversational Commerce Add-On?

Most agencies land between $500 and $4,000 per month for a conversational commerce add-on, depending on catalog size, the number of automated channels, and how much the server-side recommendation logic needs to be tuned. One-time setup fees typically run $1,500–$8,000 on top of that. If an agency quotes you a flat number without asking about your SKU count or average order value, treat it as a red flag.

Conversational commerce refers to any system where a buyer interacts with a store through a messaging interface — Instagram DMs, Messenger, WhatsApp, SMS — and receives product recommendations, answers, or a checkout link in return. A well-built implementation keeps the catalog logic on the server (live inventory, real pricing, margin rules) and lets the AI handle only the language layer.

The Three Main Pricing Models Agencies Use

1. Flat Monthly Retainer

The most common structure for mid-market stores. The agency charges a fixed fee covering setup, flow maintenance, A/B testing of message copy, and a defined number of support hours. Retainers in this category look like:

• Entry tier ($500–$1,200/month): one channel (usually Instagram or Messenger), up to 5,000 conversations/month, pre-built flows with light customization.
• Growth tier ($1,500–$2,800/month): two to three channels, custom recommendation logic, monthly reporting, and copy revisions.
• Scale tier ($3,000–$5,000+/month): full omnichannel, dedicated account manager, integration with loyalty programs or ERPs, and SLA-backed response times.

Flat retainers favor stores with predictable conversation volume. If you run a flash-sale brand with extreme seasonal spikes, a pure retainer can leave the agency absorbing costs during peaks — expect a conversation cap clause in the contract.

2. Revenue Share

Some agencies, particularly those with strong confidence in their conversion lift, propose a revenue-share arrangement: typically 3%–8% of attributed revenue generated through the conversational channel. This aligns incentives — the agency only earns more when you sell more.

The catch is attribution. Revenue share agreements require agreeing upfront on what "counts." Common rules include: the buyer clicked a link delivered via DM within a 7-day window, or the buyer's first session came from a Messenger opt-in. Without a clear attribution model, disputes are almost guaranteed at invoice time.

Revenue share tends to work best when the agency is launching a net-new channel (no historical baseline to argue over) and both sides trust the analytics stack. It is rare as the sole fee structure; most agencies pair a reduced base retainer with a smaller revenue-share kicker.

3. Usage-Based (Conversation or Message Credits)

Platforms that sell their own agency-facing tooling — including tools like SmartBrain that sit between the catalog and the messaging layer — often meter usage at the conversation or message level. Agencies pass this through to clients at a markup, typically 1.5×–2.5× the platform wholesale rate.

Usage-based pricing is the most transparent model for high-volume clients who can predict their conversation load, but it creates cost uncertainty for brands running aggressive acquisition campaigns. A single viral post that floods DMs can multiply the monthly invoice by 4× without warning. Responsible agencies build a monthly cap or overage notification into their MSA.

What Drives the Price Up (or Down)?

When you receive competing agency proposals, the variance usually comes from five variables:

• Catalog complexity: A store with 50 SKUs needs minimal recommendation logic. A store with 8,000 SKUs across multiple collections, variants, and bundles requires real server-side filtering — the kind SmartBrain handles by querying live inventory rather than relying on the AI to guess.
• Number of channels: Each platform (Instagram, WhatsApp, Messenger, SMS) has its own API quirks, opt-in rules, and rate limits. Every additional channel adds roughly 20–35% to the maintenance load.
• Personalization depth: Generic "you might also like" flows cost less than flows that factor in purchase history, browsing behavior, or loyalty tier.
• Compliance requirements: TCPA-compliant SMS programs in the US, GDPR opt-in documentation in the EU, and Meta's evolving messaging policies all add scoping hours.
• Ongoing copy and flow iteration: Static flows decay fast. Agencies that commit to monthly testing cycles charge more — and usually deliver better results over a 6-month window.

Flat Retainer vs. Revenue Share: A Quick Comparison

Neither model is universally better. The right choice depends on your margin structure and how much you trust the attribution methodology.

• Flat retainer: Predictable costs, agency effort is fixed, works well for steady-state operations and stores with thin margins where sharing revenue is painful.
• Revenue share: Variable costs, agency is incentivized to maximize conversions, works well for high-AOV stores where a few extra conversions per month easily justify the percentage.
• Hybrid (retainer + kicker): The most common structure among experienced agencies. A modest base ($800–$1,500) covers maintenance; a small share (2–4%) rewards performance. This protects the agency's floor while keeping upside tied to results.

What Agencies Should Spell Out in Every Proposal

If you are an agency building a conversational commerce practice — or a store owner evaluating one — these line items should appear explicitly in any statement of work:

• Monthly conversation or message volume included, and the per-unit overage rate.
• Which party owns the subscriber list and opt-in data if the relationship ends.
• How catalog sync works — manual CSV uploads, API polling, or real-time server queries (the last option is what separates tools like SmartBrain from simpler chatbot builders).
• SLA for flow updates when a product goes out of stock or a price changes.
• Attribution window and the analytics tool used to measure it.

Missing any of these details is how agencies and clients end up in disputes six months into an engagement.

FAQ

Is a $500/month retainer realistic for a small Shopify store?

Yes, for a single-channel implementation with a small catalog and pre-built flows. Expect limited customization and slower response times on support requests. It is a reasonable starting point to validate the channel before committing to a larger scope.

How do agencies handle platforms that charge per conversation, like WhatsApp Business API?

Meta charges per 24-hour conversation window, and rates vary by country. Agencies typically pass this cost through at cost or with a small administrative markup (10–15%), then charge their own fee separately for flow management. Ask for an itemized breakdown so you can see the platform cost versus the agency margin.

What is a reasonable revenue-share percentage?

For a mature conversational channel with proven attribution, 3–5% of attributed revenue is common. Higher percentages (6–10%) are sometimes offered in early engagements where the agency is taking on more setup risk. If you are already running a DM channel and want an agency to take it over, negotiate toward the lower end of that range.

Does the AI need to know the full product catalog?

In a well-architected system, no. The recommendation decision — which product fits this buyer's budget and needs, and whether it is in stock — should live on the server, not inside the AI model. The AI's job is to write a helpful, on-brand message around the product the server has already selected. This separation is what makes systems like SmartBrain more reliable at scale than purely prompt-driven chatbots.

What questions should I ask before signing with an agency?

Ask how the catalog stays in sync with your live inventory, who owns the subscriber data, what happens to flows when a product sells out, and how attribution is measured. Any agency that cannot answer all four clearly has not scoped the project thoroughly enough.