Marketplace agencies are usually good at saying yes. Yes, we can launch another marketplace. Yes, we can check the feed. Yes, we can add Cdiscount to the roadmap. Yes, we can include Walmart Connect in next month’s plan. Yes, we can investigate why TikTok Shop orders do not match Shopify.
That helpfulness wins clients. It can also quietly destroy agency margin.
The named mistake I see is treating every marketplace request as a delivery task before it has passed an intake desk. A client asks for “just a quick marketplace check”. The account manager forwards it to a specialist. The specialist opens three dashboards, exports sales, checks stock, reviews ads, asks operations for return data, and turns a vague question into four hours of senior work. Nobody priced it. Nobody ranked it against other clients. Nobody defined the decision the work should support.
My stance: marketplace agencies with five or more people need a marketplace intake desk. Not a ticket inbox. Not a project-management board full of half-described tasks. A commercial triage layer that decides whether a client request is support, analysis, strategy, channel launch, campaign work, catalog cleanup, or out-of-scope consulting before the team starts burning expert hours.
This guide is for agencies in Germany, the United States and cross-border teams managing Amazon, Walmart, Cdiscount, bol, Kaufland, Otto, TikTok Shop, eBay, Shopify and Mirakl retailers. Connected data matters. But agencies also need a disciplined way to decide which client questions deserve action, which deserve a paid change request, and which should wait.
What the existing marketplace software advice gets right
The research for this topic was useful. Cdiscount’s own marketplace pages explain why the channel attracts agencies and sellers: a large French audience, millions of active customers, a loyalty program, fulfillment services, advertising options and seller requirements such as complaint-rate and on-time-delivery expectations. Sellermania goes deeper into operational needs on Cdiscount: order acceptance, reimbursement rates, stock synchronization, listing creation, delivery modes and support. Those are real agency workloads, not brochure details.
ChannelEngine frames the broader software problem well: product management, inventory control, order processing, pricing, compliance and integrations. Productsup is strong on feed management, content optimization, AI mapping and access controls. Rithum talks about profitable multi-marketplace management. MerchantSpring speaks directly to agencies with portfolio oversight, client-specific reporting and scheduled reports. Pacvue emphasizes cross-retailer media, inventory signals and budget reallocation.
All of that is valuable. It also leaves a gap.
Most content assumes the work has already been approved. A product needs to be listed. A report needs to be sent. A campaign needs to be optimized. A channel needs to be integrated. But in a real marketplace agency, the expensive moment happens earlier: when a client asks an unclear question and the agency starts working before the request has a commercial owner, decision type, proof standard or scope boundary.
The unique problem for marketplace agencies: requests arrive mixed
A marketplace agency does not receive neat work packages. It receives mixed signals.
- “Can we sell on Cdiscount?” might be a strategic expansion question, a feed mapping task, a French VAT concern, a fulfillment question, or a pricing-risk conversation.
- “Why did Amazon sales drop?” might be an ad issue, stockout, Buy Box loss, price gap, suppressed listing, ranking change, return spike or competitor promotion.
- “Can you add TikTok Shop?” might be a channel launch, creator-commerce project, inventory-risk review, order-sync task, or a completely new service line.
- “Can we increase Walmart budget?” might be a retail-media opportunity, but it might also be blocked by stock cover, contribution margin or listing quality.
If those requests go straight into execution, the agency becomes reactive. Specialists start investigating in parallel. Account managers chase updates. Clients receive smart but inconsistent answers. The agency owner later wonders why a €4,500 monthly retainer needs 29 hours of work to stay calm.
The intake desk fixes that by adding one step before execution: classify the decision.
The intake desk should classify requests into seven decision types
I like seven buckets because they are specific enough to change routing, but simple enough for account teams to use without a workshop.
1. Support request
This is a small operational fix inside the agreed scope: a missing report, a dashboard access issue, a known marketplace error, a client user who needs a permission change. It should have a short SLA and a clear owner. It should not invite a strategic investigation.
2. Performance explanation
This request asks why a number moved: sales, ACOS, TACoS, conversion, stockouts, returns, Buy Box, content score or marketplace ranking. It needs evidence, not just action. FiveX fits naturally here because the team can reconcile marketplace analytics, ads, profitability, inventory and product groups before replying with one explanation instead of five exports.
3. Spend permission
This request moves money: higher bids, more daily budget, new retail-media campaigns, extra Cdiscount Advertising activity, Walmart Connect scaling or Sponsored Products expansion. It should require margin, stock and target evidence before anyone touches budget. In FiveX, agencies can connect ad performance with SKU-level profitability and inventory signals, so “scale” becomes a permissioned decision rather than a confident guess.
4. Catalog or feed change
This includes category mapping, attributes, titles, variants, product groups, marketplace taxonomy, images and localization. These changes look operational, but they can affect search visibility, ad eligibility and reporting structure. Productsup and ChannelEngine are strong on the feed execution layer; the agency still needs a profit-aware approval layer around which changes are safe, risky or client-approved.
5. Channel launch
This is any request to add or expand a marketplace: Cdiscount, Kaufland, Otto, Walmart, Target Plus, TikTok Shop, Mirakl retailers or a country variation. It should never be routed as “please connect the channel” first. It needs a launch score: SKU margin, fees, fulfillment fit, support language, stock runway, ad plan, client owner and first-90-day success metric.
6. Change request
This is work outside the retainer: a new weekly report, custom dashboard, one-off marketplace audit, new country, new product group structure, extra QBR analysis, retail-media rebuild or emergency cleanup after another vendor made changes. If the intake desk cannot identify paid change requests, the agency will donate its best thinking for free.
7. Executive decision
Some requests should go straight to leadership: pausing an unprofitable channel, changing the client’s service tier, refusing an expansion, renegotiating scope, or challenging the client’s margin assumptions. A good intake desk protects specialists from carrying decisions that belong in the commercial relationship.
Named example 1: the Cdiscount “quick check” that was not quick
Imagine a German marketplace agency with 8 employees managing a home-and-living brand across Amazon.de, bol.com and Shopify. The client asks, “Can we also try Cdiscount? It looks big in France.”
Without an intake desk, the account manager asks the marketplace specialist to investigate. The specialist checks Cdiscount requirements, looks at 320 SKUs, asks the feed manager about French attributes, asks operations about delivery options, checks rough commission assumptions, and asks the ad specialist whether launch budget is needed. By Friday, the team has spent 7.5 hours. At an internal blended cost of €68 per hour, that is €510 of agency cost before the client has approved anything.
The answer is also incomplete. Nobody has checked French support capacity, on-time delivery risk, fulfillment impact or whether the client accepts contribution margin instead of GMV as the first-90-day target.
With an intake desk, the same request becomes a channel-launch pre-scope:
- Decision type: channel launch.
- Initial cap: 90 minutes unpaid discovery, then paid launch assessment.
- Evidence required: top 50 SKU margin, expected marketplace fees, fulfillment route, French support plan, stock cover, launch budget and owner on client side.
- Commercial rule: no catalog work starts until the client approves a €1,800 launch assessment or trades it against existing roadmap work.
That makes the help visible, priced and useful.
Named example 2: the Amazon sales drop that should not become an ad rebuild
Now take a US agency managing a consumer electronics client on Amazon, Walmart and Shopify. Monday morning, the client says Amazon revenue fell 22% week over week and asks the team to “fix the ads”.
A reactive team opens the ad platform first. They lower bids in two campaigns, add negatives, and rebuild one Sponsored Products structure. The ad specialist spends three hours. The account manager spends another hour explaining the changes. Two days later, revenue is still soft.
The intake desk would classify this as a performance explanation before spend permission. The first question is not “which bid should move?” It is “which constraint moved?”
The team checks FiveX and sees:
- Amazon ad spend was down only 4%, not 22%.
- The hero SKU had 9 days of stock left, so the buy box share became unstable.
- Contribution margin dropped from 18% to 11% because a coupon overlapped with higher referral and fulfillment costs.
- Walmart revenue rose 14% on the same SKU family with cleaner margin.
- Return rate on one bundle moved from 6% to 13% after a content change.
The recommendation changes completely: freeze Amazon bid increases, protect stock, remove the coupon overlap, route demand to the healthier Walmart bundle and open a catalog QA task. The agency did not “fix ads”. It protected profit.
Named example 3: the retainer that looked profitable until request load was counted
A marketplace agency has a €6,000 monthly retainer with a beauty brand. On paper, it looks healthy. The planned scope is 42 hours per month, so at €95 blended billing value per hour the account has room for margin.
But the intake desk shows a different picture. In one month, the client submits 38 requests:
- 11 support requests
- 9 performance explanations
- 6 spend-permission questions
- 5 catalog changes
- 3 channel-launch ideas
- 3 custom reporting requests
- 1 executive escalation
The team spends 64 hours, not 42. At €61 internal cost per hour, delivery cost is €3,904 before management time. The retainer is still cash-positive, but the senior team is subsidizing indecision. The real issue is not “the client is demanding”. It is that the agency has no request rules.
A better model sets thresholds: support stays inside scope, performance explanations are batched twice per week, spend decisions require FiveX evidence, catalog changes go into a weekly release, and channel launches become paid assessments.
What good intake software needs to show
A marketplace intake desk can start as a disciplined process, but it becomes much stronger when software supports the evidence. The tool stack should answer six questions quickly:
- Which client, channel and SKU group does this request affect? Portfolio context matters. A €500 ad decision for a small account is not the same as a €500 ad decision for a launch SKU with low stock.
- Is the request inside scope? The account team needs a visible service boundary, not a memory test.
- Which metric moved, and is it fresh? Sales, profit, ad spend, ACOS, inventory, returns and conversion should sit close enough that the team does not build a spreadsheet for every answer.
- Who owns the decision? Support, ads, catalog, operations, account lead and client owner should not be guessed in Slack.
- What proof is needed before action? Spend changes need margin and stock. Channel launches need launch scores. Catalog changes need visibility and reporting impact.
- What happens if we do nothing this week? Not every request deserves immediate action. Good intake shows risk, not just urgency.
This is where FiveX should sit in the agency operating model: not replacing feed tools, ad platforms or project management, but acting as the profit and evidence layer behind the intake decision. The product hooks are practical: multi-channel analytics to explain movement across marketplaces, advertising automation and bid recommendations with guardrails, SKU-level profitability and margin data, inventory risk signals, product groups, client-ready reporting and AI agents that can investigate recurring questions from live marketplace data.
The operator rule: never accept a vague request into expert time
Here is the rule I would put on the wall: no vague request enters expert time unchanged.
Before a specialist works on it, the account lead or intake owner should translate the request into a decision sentence:
- “Decide whether Cdiscount deserves a paid launch assessment for the top 50 SKUs.”
- “Explain whether Amazon revenue dropped because of ads, stock, price, content, returns or marketplace mix.”
- “Approve or reject a €75 daily Walmart Connect budget increase based on margin and stock.”
- “Classify this reporting request as inside retainer, paid change request or QBR backlog.”
That small rewrite changes the whole agency. Specialists stop becoming detectives for unclear questions. Account managers stop forwarding anxiety. Clients get stronger answers. Leadership can see which accounts create profitable strategic work and which ones create unmanaged request load.
A simple marketplace intake desk template
Use this structure for every incoming request:
- Client: account name and service tier.
- Request sentence: what the client asked, in their words.
- Decision sentence: what the agency must decide.
- Decision type: support, performance explanation, spend permission, catalog/feed change, channel launch, change request or executive decision.
- Commercial status: inside scope, capped discovery, paid change request, roadmap backlog or leadership review.
- Evidence required: sales, ad spend, profit, stock, returns, listing quality, client approval, fee assumption or launch score.
- Owner: one person, not a channel.
- Time cap: how much unpaid work may happen before escalation.
- Response format: Slack answer, client email, dashboard note, QBR slide, task, or proposal.
The time cap is important. A 30-minute cap for support, 90-minute cap for first analysis and zero unpaid execution for new channel launches sounds strict. It is actually generous. It lets the agency be helpful while protecting the team from invisible over-service.
How FiveX helps agencies run this without another spreadsheet
FiveX helps when the intake desk needs evidence fast. Account managers can check revenue, ad spend, profit, returns and inventory risk before assigning work. Ad specialists can see whether a bid change is safe against margin and stock. AI recommendations and automation rules can support the next action, while the intake desk decides whether money is allowed to move.
The most useful output is not another dashboard screenshot. It is a better client sentence:
“We checked the last 14 days. The sales drop is not primarily an ad issue: spend is stable, but stock cover on the hero SKU is down to 9 days and contribution margin fell to 11% after the coupon overlap. We recommend pausing bid increases, fixing the promotion overlap, and moving the Walmart bundle into this week’s growth slot. A full Cdiscount launch assessment would be a separate scoped project.”
That sentence is calm, commercial and defensible. It turns software into agency margin.
Final thought
Marketplace agency software should not only make execution faster. Faster execution without intake control just helps an agency lose margin at a higher speed.
The better goal is decision discipline. Every client request should become a classified decision before it becomes specialist work. Every spend change should earn permission from profit, stock and context. Every channel-launch idea should be scoped before catalog work starts. Every custom report should be tied to a service boundary.
That is the role of the marketplace intake desk: protect the team, sharpen the client conversation and turn FiveX into the proof system behind profitable agency work.