E-commerce manager: running the website’s revenue day to day, with reliable data and checks that run on their own.
The e-commerce manager is accountable for online revenue: tracking traffic and conversion rate, managing service providers and agencies, preparing campaigns and fixing what blocks. The day often starts with exports to cross-check between the e-commerce platform, the web analytics tool and the OMS. Automation gives that time back for decisions.
The role
Between management, merchandising, marketing, logistics and technical service providers.
The e-commerce manager runs the website like a store: making sure products are online, the checkout works, campaigns start on time and orders go out. They prioritise the changes requested from service providers, track acquisition costs with marketing and report results to management. In a mid-sized company, they often combine the roles of website project manager and performance lead.
Day to day
A typical day.
- MorningRead yesterday’s sales, traffic, conversion rate and average basket, and understand the variances.
- MorningCheck the website is running: checkout, payment, pages in error, products with no price or no visual.
- MiddayCatch up with merchandising and marketing on current and upcoming campaigns.
- AfternoonFollow up change requests with the agency or the integrator, and test what is going into production.
- AfternoonSort out blocked orders, feedback from customer service and questions from logistics.
- End of weekPrepare the management report: revenue against target, acquisition, returns, margin.
The objectives
What the role is measured on.
- Online revenue against target
- Conversion rate
- Average basket
- Acquisition cost by channel
- Return rate
- Website availability
Indicators are named, never quantified: levels depend on each company, and the diagnostic measures them.
Day to day
The problems that come up most.
Figures that do not match
The platform, the web analytics tool, the OMS and finance each give a different revenue figure, and the gap has to be explained before the result can be discussed.
Faults discovered by customers
A payment method down, an empty product page or a promo code that does not work are often reported by customer service, after hours of lost sales.
Testing done by hand
Every release means going through the purchase journey again on several devices and in several countries, and regressions slip through.
Reporting time that eats into analysis
Monday goes on exports and spreadsheets; little time is left to understand why conversion dropped.
Service providers that are hard to manage
Change requests, tickets and deliveries are tracked in e-mails and spreadsheets, with no clear view of what is expected, delivered and tested.
Campaigns that overlap
Sales, private sales, launches and promo codes overlap; stacking rules and dates are checked by hand.
By sector
One role, several realities.
- Fashion and luxury
A website that must reflect the brand’s standards, smaller volumes of high value, services such as in-store appointment booking or gift wrapping, and particular attention to fraud on high-value orders.
- Retail
A website backed by a store network: click & collect, ship-from-store, in-store returns, attribution of online sales to stores and campaigns synchronised with the network.
- Cosmetics and beauty
Frequent launches, gift sets and limited editions, samples added to orders, many customer reviews to moderate and campaigns shared with retailers and marketplaces.
- B2B distribution and wholesale
A wide and changing catalogue, prices and stock updated often, a share of trade customers with their own terms, and a website that lives alongside several marketplaces.
What changes
Automate the task, keep the judgement.
What gets automated
- Daily reporting of sales, traffic, conversion rate and average basket, reconciled between the platform, the web analytics tool and the OMS.
- Continuous website monitoring: purchase journey, payment, pages in error, products with no price or no visual, with an alert to the right person.
- Automated tests of the purchase journey before each release, on the main devices and countries.
- Detection of conversion variances by device, traffic source and checkout step, with a first diagnosis drafted by an AI agent.
- Tracking of promo codes, orders at risk of fraud and customer reviews, with simple cases handled without intervention.
What stays human
- Website development priorities and budget trade-offs with service providers.
- Commercial decisions: campaigns, discounts, product highlighting, with merchandising and marketing.
- Validation of suspicious orders and replies to sensitive customer reviews.
Use cases
How to automate, case by case.
Each case starts from a question leadership teams ask. The starting point, what gets automated, the tools involved, and what stays in the teams’ hands.
How do you automate daily e-commerce reporting?
- Today
- Every morning, the e-commerce manager exports sales from the platform, traffic from the web analytics tool and shipments from the OMS, then assembles them in a spreadsheet. Yesterday’s figures arrive late in the morning, and the gaps between them still need explaining.
- What gets automated
- Data from the three tools flows automatically every night into a single database, with definitions fixed once: orders, gross and net revenue, sessions, conversion rate, average basket, share of new customers. The dashboard is ready first thing, compared with the previous day, the same period last year and the target. An AI agent drafts a short commentary on the notable variances.
- What stays human
- The e-commerce manager interprets the variances and decides what to report to management.
- E-commerce platform
- Web analytics tool
- OMS
- Reporting tool (BI)
How do you automate the analysis of a drop in conversion rate?
- Today
- The conversion rate has been falling for a few days; the e-commerce manager slices the data by hand by device, traffic source and page, without knowing where to start. The analysis takes hours and often lands after the problem has ended.
- What gets automated
- Conversion rate is tracked continuously by device, browser, country, traffic source and checkout step. When a segment drops, the tool isolates the step where visitors abandon and links it to current events: a release, stockouts on highly viewed products, a payment incident, a campaign change. An AI agent presents the hypotheses ranked by likelihood.
- What stays human
- The e-commerce manager checks the chosen hypothesis and decides on the fix with the service provider or merchandising.
- Web analytics tool
- E-commerce platform
- Reporting tool (BI)
- Messaging
How do you automate testing of an e-commerce website before a release?
- Today
- Before each agency delivery, the team goes through the purchase journey by hand on desktop and mobile, in a few countries. For lack of time, some cases are not tested, and a regression on a payment method or a promo code goes into production.
- What gets automated
- Critical journeys are written once as test scenarios: search, product page, add to basket, applying a promo code, choosing the delivery method, payment, account creation. They run automatically on the test environment at each delivery, on several devices and countries, and the result is attached to the ticket. The release is blocked while a critical scenario fails.
- What stays human
- The team tests new features and visual rendering itself, and decides on the release.
- Automated testing tool
- E-commerce platform
- Ticket tracking tool
How do you automate the detection of outages and errors on an e-commerce website?
- Today
- A payment method rejects cards, a category page shows an error or the website slows down in the evening: the team finds out from an unhappy customer or from the next day’s sales. Nothing links the incident to the time it started.
- What gets automated
- Automatic monitoring continuously places test orders on the website, checks key pages, measures loading time and compares order volume by payment method with what is expected. Each anomaly triggers an alert to the e-commerce manager and the service provider concerned, with the start time and the pages and countries affected.
- What stays human
- The e-commerce manager assesses severity, decides on any communication to customers and follows the fix with the service provider.
- Website monitoring tool
- E-commerce platform
- Messaging
How do you automate tracking of acquisition spend and its return by channel?
- Today
- Spend on online advertising, affiliation and comparison sites is collected from each ad platform, then matched by hand with sales from the web analytics tool. The real acquisition cost, net of returns, is only known at month end.
- What gets automated
- Spend from each ad platform and attributed sales flow every day into a single database, with returns and cancellations from the OMS. Acquisition cost and return by channel, campaign and product family are calculated with the same attribution rule, and campaigns that spend without selling are flagged.
- What stays human
- Budget allocation between channels is decided by the e-commerce manager and marketing.
- Ad platforms
- Web analytics tool
- OMS
- Reporting tool (BI)
How do you automate tracking of online revenue against target?
- Today
- The annual target is split by month in a spreadsheet; progress is recalculated by hand every week, without taking the sales calendar or campaigns into account. Management discovers the shortfall at month end.
- What gets automated
- The target is spread by day according to last year’s sales profile and the calendar of planned campaigns. Every day, actuals are compared with this curve, and a month-end projection updates. Variances are broken down into traffic, conversion rate and average basket.
- What stays human
- The e-commerce manager proposes catch-up actions; management decides on discounts or additional budgets.
- E-commerce platform
- OMS
- Reporting tool (BI)
- Spreadsheet
How do you automate moderation of and replies to customer reviews?
- Today
- Reviews arrive on the website and on the review platform; someone reads them one by one, publishes or rejects, and replies when there is time. Reviews reporting a product defect or a delivery problem do not reach the teams concerned.
- What gets automated
- An AI agent reads each review, checks the publication rules, classifies the topic (size, quality, delivery, service) and drafts a reply in the brand’s tone. Positive reviews with no question are published according to the rules, negative reviews reach the team with the related order, and recurring comments on a product are passed to product and merchandising.
- What stays human
- The team validates replies to negative reviews and decides on any goodwill gesture.
- Customer review platform
- E-commerce platform
- OMS
- Customer service tool
How do you automate opening an e-commerce website in a new country?
- Today
- To open a country, the team translates product pages and content by hand, copies prices into a new currency, sets up carriers and legal notices, then checks product by product. The opening slips and omissions appear after launch.
- What gets automated
- The list of products allowed in the country, their prices and their translations is prepared from the PIM and the price source, with translation by an AI agent and review of sensitive content. An automatic checklist verifies, before opening, that every published product has a price, a translation, visuals and a delivery method, and that legal pages and local payment methods are in place.
- What stays human
- The choice of assortment, prices by country and terms of sale stays with management, with legal for local statements.
- PIM
- E-commerce platform
- Translation tool
- Carrier platform
How do you automate promo code management?
- Today
- Promo codes are created by hand in the platform for each campaign, partner or goodwill gesture. Stacking rules with sales are checked from memory, codes stay active after the campaign and circulate on deal websites.
- What gets automated
- Each code is created from a campaign sheet with its rules: dates, excluded products, stacking allowed or not, single or multiple use. Before activation, a check simulates stacking with current promotions, and monitoring detects codes used outside their target or after their end date. Expired codes are deactivated automatically.
- What stays human
- The e-commerce manager and marketing decide on the codes, their value and whether to stop a misused code.
- E-commerce platform
- OMS
- Reporting tool (BI)
How do you automate the review of orders suspected of fraud?
- Today
- Orders flagged by the payment provider’s anti-fraud tool wait in a back office, reviewed by hand when someone has time. Good orders ship late, and some fraud still gets through.
- What gets automated
- Risk signals are gathered on each order: payment provider score, mismatch between billing and delivery address, customer history, high-value products, device used. Low-risk orders are released automatically, high-risk orders are held with the file prepared, and confirmed fraud feeds back into the rules.
- What stays human
- The team decides whether to ship or cancel held orders, and adjusts the accepted level of risk.
- Payment provider
- OMS
- E-commerce platform
- Customer service tool
Questions from the teams
What teams look to improve.
- How do you automate performance tracking for an e-commerce website?
- By first fixing the definitions: gross or net revenue, validated or shipped orders, sessions or visitors. Data from the platform, the web analytics tool and the OMS then flows every day into a single database, and the dashboard is always calculated the same way. The e-commerce manager stops cross-checking and starts analysing.
- Why does revenue differ between the platform and the web analytics tool?
- Because they do not measure the same thing: the web analytics tool depends on cookie consent and blockers, the platform counts orders placed, the OMS orders shipped, finance orders paid. Automation documents each gap and picks one reference source per indicator.
- How do you get alerted when the website is losing sales?
- By continuously comparing order volume and conversion rate with what is expected for the hour and the day, and by regularly testing the purchase journey end to end. An abnormal drop triggers an alert with the first clues: payment method, device, country, page concerned.
- Can e-commerce website testing be automated?
- Largely. Critical journeys (search, add to basket, payment, account creation, promo codes) run automatically before each release. Human testing focuses on new features and visual rendering.
- Can AI explain a drop in conversion?
- It can isolate where the drop happens and link it to known events: a release, stockouts on key products, a change in traffic, a payment incident. It proposes ranked hypotheses; the e-commerce manager checks them and decides on the action.
- How do you manage an e-commerce agency or integrator more simply?
- By bringing change requests, tickets, estimates and deliveries together in a single tracking tool, linked to automated testing. Each delivery is tied to its request and its test result, which makes budget trade-offs easier.
- Do you need a reporting tool (BI) for e-commerce reporting?
- Not necessarily at first: an automatically fed dashboard is often enough. A reporting tool becomes useful when the website has to be combined with stores, marketplaces, stock and margin.
- Which e-commerce manager tasks are automated first?
- Daily reporting, website monitoring and purchase journey testing: they come back every day, are easy to describe and their errors cost sales directly.
02 — Getting started
Describe what’s holding your teams back.
A first conversation to qualify your need. Then a scoped diagnostic: the scope, the solution, the estimate.
Discuss your projectReply within 48 business hours.