How was ROAS measured for this ecommerce company?

ROAS used shipped and confirmed sales from SAP connected to Magento, UTMs, the CRM and a SQL database. Cancelled, refunded and returned orders were removed before return was calculated.

What did the AI handle in the sales process?

Chat and voice agents responded within seconds, called the database through webhooks to match vehicles with tires and parts, generated the payment path, and handed conversations to a person when needed.

Can this system be copied for any auto-parts store?

The measurement and decision methodology transfers, but the implementation is not a template. It must fit the catalog, ERP, ecommerce platform, operations and sales-team availability.

Why is the client not named?

The client remains anonymous because its commercial and operational data is confidential. Adsformance publishes only the process and outcomes authorized for public use.

Auto Parts Ecommerce Case Study: 10-15x ROAS with AI

10-15x

monthly blended ROAS

Up to 30x during Cyber periods

7,200+

sales conversations/month

WhatsApp and Messenger

40-60%

year-over-year growth

Reported sales value

5 years

working relationship

Continuous modernization

Source: operating data confirmed by the client and Adsformance across Magento, SAP, CRM and ad reporting. The client remains anonymous for confidentiality. Results vary with the market, offer, operations and data quality.

By Adsformance · Operator interview with Furqan · Updated 2026-09-23

Growth did not come from one campaign. It came from turning a legacy, fragmented operation into a system where acquisition, catalog, customer service, payment, fulfillment and attribution share data. That integration supports 10-15x monthly blended ROAS, more than 7,200 sales conversations per month and customer responses within seconds.

The challenge: several agencies, no owner of the complete result

Before Adsformance, the company had tried several agencies. One handled design, another ran media, and no partner owned the full customer journey. There was no consistent design language, tracking was broken and the owners lacked a reliable view of return.

The company also had an operating reality many digital proposals overlook: it was a long-established business running Magento with SAP and commercial processes created before modern ecommerce. Modernization required understanding that operation, proving value with data and bringing the owners along without disconnecting the existing business.

The actual assignment was not “manage ads.” It was to design, measure and operate the digital transformation across the website, creative, media, sales, CRM, service, attribution and reporting.

How ROAS was verified without trusting platform reporting

Magento recorded created orders, but not every order became final revenue. Adsformance connected ecommerce and SAP data through APIs, a SQL database, UTM parameters and the CRM. Only confirmed and shipped SAP sales were used for ROAS.

Data source	Role in measurement
Magento	Order, product, payment link and ecommerce behavior.
SAP	Final confirmation that a sale was completed and shipped.
SQL + APIs	Connected product, vehicle, UTM, order and revenue data.
CRM	Conversation, reactivation, service and contact status.
Meta and Google	Media cost and feedback from verified conversions.

Cancelled, refunded and returned orders were removed. Product data and margin could be reviewed before spend decisions, instead of optimizing against gross value reported by an isolated platform.

“The methodology can transfer; the implementation is not a template. The advantage comes from integrating the operation and making the right decisions quickly.”

Revenue architecture

From the ad to the order confirmed in SAP

The system connects acquisition, product matching, conversation, payment, fulfillment, reactivation and media learning.

01

Meta and Google

Capture demand and route traffic to Magento, WhatsApp or Messenger based on intent.

02

Chat and voice agents

Respond within seconds, understand the request and qualify the need before the close.

03

Webhook matching

The request calls SQL and returns tires or parts compatible with the customer's vehicle.

04

Payment and fulfillment

Magento creates the payment link; SAP confirms when the order is completed and shipped.

05

Human handoff

If a customer requests help or the conversation goes off track, the CRM assigns an available rep by schedule.

06

Attribution and reactivation

The confirmed sale returns to the CRM, Meta and Google. The CRM later activates repurchase follow-up.

Operating and commercial results

10x to 15x monthly blended ROAS, with low-season months closer to the bottom of the range.
Up to 30x ROAS during Cyber periods, supported by a high average ticket.
More than 7,200 sales conversations per month initiated through media and messaging.
40% to 60% year-over-year growth in sales value during the working relationship, based on shared commercial data.
Initial response within seconds and roughly 90 seconds from start to close for a typical automated conversation.
Routine call-center work handled by approximately three people, compared with about ten before message handling and qualification were automated.

Automation did not remove human judgment. It reduced repetitive conversations, accelerated product matching and reserved the sales team for cases requiring judgment, negotiation or special assistance.

The lesson: repeat the method, design the system

Another auto-parts ecommerce company can adopt the same discipline: use final sales as the source of truth, connect media to commercial systems, return verified conversions and respond quickly. It should not copy this architecture without reviewing its catalog, ERP, ecommerce platform, compatibility rules, fulfillment capacity and sales schedules.

Adsformance's advantage was bringing design, paid media, data, CRM, automation and operations into one team that could make decisions without waiting for several agencies to coordinate. Five years of continuity matters because the system kept changing with the business instead of ending at the first automation launch.

This approach is a poor fit when product data is disorganized, inventory is unreliable, no final sales source exists, the business will not change its processes, or AI is expected to compensate for a weak offer.

Related services

Ecommerce performance marketingPaid media, CRO, retention, attribution and margin-aware growth. CRM and AI-agent implementationAutomation, reactivation, voice, messaging and sales handoff. How to evaluate an ecommerce agencyReal ROAS, margin, attribution, creative and retention.

Frequently asked questions

Using shipped and confirmed SAP sales connected to Magento, UTMs, the CRM and SQL. Cancelled, refunded and returned orders were removed.

It responded through chat or voice, queried the compatibility database through webhooks, recommended products, created the path to payment and handed the case to a person when required.

The methodology can adapt. The architecture must be designed for each company's catalog, ERP, ecommerce platform, fulfillment and sales process.

Its commercial and operating data is confidential. Only the process and outcomes authorized for public use are included.

Ecommerce, data and AI

Does platform ROAS match the orders you actually shipped?

We review your ecommerce platform, ERP, tracking, campaigns and follow-up to find where margin or attribution is being lost.

Send your case See more results