Search terms are where marketplace advertising accounts tell the truth.
Campaign names can be clean. ROAS can look tidy. A monthly report can say “efficiency improved” and still hide the fact that the account is relearning the same queries every quarter. The real memory of an Amazon, bol or MediaMarkt account lives in the search terms: which shopper phrases deserve exact control, which ones deserve discovery budget, which ones should be blocked, and which ones only work when margin and stock are healthy.
The operator mistake I see in agency handovers is keyword amnesia. A new Advertentie Service partner takes over, exports the active campaigns, rebuilds the structure, maybe imports the obvious exact keywords, and starts optimizing. It feels professional. But the old account knew things the export did not explain: why “wireless charging pad fast” was blocked on Amazon, why a Dutch “hoesje iPhone 15 Pro Max” query was allowed to keep spending on bol even at 31% ACOS, why MediaMarkt category traffic was capped until the bundle had 45 days of stock. When that memory disappears, the next €5K of spend pays tuition fees.
My stance is simple: a search term handover should be treated as a profit memory transfer, not as a keyword export. The receiving operator should inherit three ledgers: winners, blockers and watchlist terms. Each term needs an owner, a reason and a permission rule. Without that, the new team is not taking over an ad account. It is taking over a guessing machine.
This guide is written for ecommerce brands in the Netherlands and Belgium spending from roughly €5K per month across Amazon, bol and MediaMarkt. At that level, search term discipline is no longer “PPC hygiene”. It is one of the cheapest ways to stop repeated waste, protect margin and make agency transitions less dramatic.
What the market already explains well
The competitor content around search terms is strong on mechanics. BidX explains why Amazon campaigns need structure: budget is controlled at campaign level, so auto, broad, exact and product targeting often deserve separation. Their guidance on moving profitable terms through match types is useful because it gives operators cleaner bid control.
BidX also explains Amazon Advertising reports clearly. The Sponsored Products search term report shows which shopper queries or competitor ASINs led to clicks and sales. The targeting, advertised product, placement and purchased product reports add context. That is the right starting point: you cannot optimize what you cannot see.
SellerMetrics and Pattern both cover negative keywords well. They explain negative exact versus negative phrase, the value of blocking irrelevant or unprofitable queries, and the idea of search term isolation: when a winning term moves from auto or broad into manual exact, the discovery campaign should often receive a negative so the same traffic is not bought twice.
Podean’s Sponsored Products guide gives the broader auction context: Amazon filters for relevance and offer quality before sponsored ads compete. bol’s own partner platform explains the same commercial promise for Sponsored Products: ads connect shopper search terms to items, improve visibility, and can support organic ranking when the product is eligible and relevant. MediaMarktSaturn’s retail media materials are more format-led, but the same logic applies: sponsored search and native product ads are only as good as the product, offer and category context behind the click.
What most of this advice misses is the handover problem. It tells you how to find, harvest and negate terms inside one account at one moment in time. It rarely tells you how to preserve the commercial reason behind those decisions when a new agency, freelancer, internal specialist or automation layer takes over. That is where expensive relearning starts.
The search term handover ledger
A useful handover is not a spreadsheet with thousands of exported rows. Nobody needs a museum of every query that ever received one impression. The goal is to transfer decision memory.
I like a three-ledger structure:
- Winner ledger: search terms that have earned controlled budget because they convert, protect profitable demand or open an attractive growth lane.
- Blocker ledger: search terms, stems or competitor targets that should be excluded because they waste spend, sell the wrong product, create margin loss or attract the wrong shopper.
- Watchlist ledger: terms that are not clearly good or bad yet, but deserve measured discovery because the economics depend on price, stock, reviews, seasonality or marketplace role.
Each row should answer five questions: what is the term, where did it appear, what decision was made, why was that decision made, and what would change the decision later?
That last question is the one most handovers skip. A negative keyword is not always permanent truth. A query may be blocked because the current SKU has 12 days of stock, because the margin is too thin at today’s price, or because the product page has weak reviews. If those conditions change, the decision might change too. FiveX helps here because marketplace analytics, product profitability and inventory insights can sit next to ad data instead of living in separate exports. The search term decision becomes visible as a commercial rule, not a mysterious PPC artefact.
Named example 1: Amazon branded defence that should not be “optimized away”
Imagine a Dutch kitchen brand spending €3,200 per month on Amazon.nl Sponsored Products. The old agency has a branded exact campaign for “Nordvik messenset” and close variations. It spends €420 per month at 11% ACOS, with 42% contribution margin and 68 days of FBA stock. A new operator sees that organic rank is already position one and suggests cutting branded spend to free budget for generic discovery.
That sounds logical until the search term ledger shows the reason behind the campaign. Two resellers also advertise on the brand term. When the branded campaign was paused for nine days in April, paid spend dropped by €126, but reseller share on the product detail page increased and the brand lost an estimated €1,040 in contribution profit on its hero bundle. The term is not just a cheap conversion keyword. It is a defence rule.
The handover row should read: keep branded exact active while contribution margin is above 35%, stock cover is above 30 days and reseller activity is visible. Do not judge this term only by incremental-looking ROAS. Judge it by defended margin and offer control.
That is exactly where a FiveX Advertentie Service workflow earns its keep. The decision is not “branded terms always deserve spend”. The decision is “this branded term deserves spend under these SKU economics and competitive conditions”. Much less catchy. Much more useful.
Named example 2: bol discovery term that looks bad until you separate Belgium from the Netherlands
Now take a home-storage seller on bol. The broad Sponsored Products campaign shows the query “opbergbox 45 liter transparant”. Across NL and BE together, the term spent €610 in four weeks, drove €1,870 revenue and sat at 32.6% ACOS. The SKU has 24% contribution margin before ads, so a quick audit would probably mark it as too expensive.
The handover ledger tells a better story. In the Netherlands, the term spent €470, converted at 7.8%, and produced only €1.90 contribution profit per order after ads. In Belgium, it spent €140, converted at 13.5%, had lower return pressure, and produced €5.80 contribution profit per order after ads. The old operator had planned to split the term: negative exact in the Dutch learn campaign, controlled exact in the Belgian scale campaign, and a lower bid in the shared broad campaign until stock improved.
If that context is not transferred, the new agency may either cut the term everywhere or keep overspending in both markets. Both decisions are lazy. The correct handover is more specific: the search term is not good or bad; it is marketplace-specific.
FiveX’s product profitability view and marketplace-level reporting make this kind of split easier to defend in a client meeting. Instead of saying “Belgium performs better”, the operator can show retained contribution margin, stock cover and return rate by SKU and marketplace. That turns the search term from an opinion into a budget rule.
Named example 3: MediaMarkt launch query that needs a sunset date
For electronics brands, MediaMarkt retail media often behaves differently from Amazon or bol. The search volume may be more category-led, the consideration cycle can be longer, and the product page has to compete inside a retailer environment where brand trust is high but comparison shopping is ruthless.
Suppose an accessory brand launches a USB-C docking station on MediaMarkt.nl with €900 launch budget. The query cluster around “usb c docking station laptop” spends €280 in week one, creates 96 clicks, 7 orders and 2.4 ROAS. At first glance, that is not enough. The SKU has 38% contribution margin and a €64.95 selling price, but after retail media spend the first week is barely profitable.
The old agency kept it on a watchlist rather than cutting it. Why? The launch target was not immediate profit alone. The product had 72 days of stock, a strong price position versus two competitors, and only 4 reviews. The term was allowed a two-week learning window capped at €650 total spend. If ROAS did not pass 3.2 or review count did not pass 12 by the end of the second week, budget would move back to Amazon exact terms with proven demand.
That is a proper watchlist rule: test budget, success condition, sunset date and fallback destination. Without it, a new team can misread the term either as a failure to pause immediately or as a launch bet to keep alive forever. Both miss the point. Launch search terms need an expiration date. Otherwise “learning” becomes a very polite word for leakage.
How to run the handover before a new operator touches bids
The handover should happen before campaign restructuring. I know that is annoying. Everyone wants to clean the account first. But if you restructure before preserving search term memory, you destroy the map and then complain that the territory is confusing.
Use this sequence:
- Export the last 90 days of search term data for Amazon Sponsored Products, Sponsored Brands where relevant, bol Sponsored Products and any MediaMarkt retail media reporting available.
- Flag material terms only: meaningful spend, meaningful clicks, meaningful sales, or strategic importance such as branded defence and competitor conquesting.
- Add SKU economics: contribution margin, return rate, fulfilment model, stock cover, Buy Box or offer eligibility, price position and marketplace role.
- Assign each term a ledger status: winner, blocker or watchlist.
- Write the reason in plain language: “block because it sells low-margin spare part”, “keep because it defends branded demand”, “test until stock drops below 25 days”.
- Translate decisions into account actions: exact keyword, negative exact, negative phrase, bid cap, campaign move, budget cap or review date.
The plain-language reason matters. A future operator may not agree with the decision, and that is fine. At least they know what they are changing. FiveX advertising automation and AI recommendations can then use the same commercial inputs: not just “ACOS is high”, but “this term is allowed to run because it protects a high-margin SKU with enough stock” or “this term is blocked because it sells a variation below break-even”.
The five handover mistakes that create repeated waste
First, exporting keywords but not search terms. Keywords are what you targeted. Search terms are what shoppers actually typed. The difference is where waste hides.
Second, transferring negatives without rationale. A negative phrase can block useful future demand if nobody remembers why it was added. “Cheap” may be a sensible blocker for a premium device. It may be a mistake for an outlet campaign.
Third, treating Amazon logic as bol logic. Amazon match types, bol Sponsored Products and MediaMarkT retail media reporting do not expose the same controls in the same way. The commercial ledger can be consistent, but platform actions must be translated.
Fourth, ignoring stock. A term that works with 90 days of stock can be dangerous with 11 days. Search term quality is not independent from inventory.
Fifth, rebuilding before reading. A beautiful new campaign structure is not a strategy if it deletes hard-earned lessons from the last operator.
What a good Advertentie Service partner should deliver
If you are hiring or replacing a marketplace advertising partner, ask for more than “account access” and “campaign exports”. Ask for the search term ledger.
A good partner should be able to show which terms were promoted, which were blocked, which are still being tested, and what commercial condition sits behind each decision. They should connect those terms to SKU margin, stock cover, marketplace role and campaign purpose. They should also be comfortable saying, “we do not know yet” and placing a term on a watchlist with a capped budget rather than pretending every query has an immediate answer.
That is how FiveX thinks about Advertentie Service for Amazon, bol and MediaMarkt accounts from €5K spend. The platform gives the operator one view of ad performance, product profitability, inventory and marketplace reporting. The service layer turns that view into weekly decisions: promote, block, cap, test, split by marketplace or pause until the product earns permission again.
The outcome is not a bigger keyword list. The outcome is a calmer account where the same mistakes do not get repurchased every month.
Final thought: protect the lessons you already paid for
Every search term report contains tuition. Some of it was cheap. Some of it was painfully expensive. If your next agency, tool or internal hire starts without that memory, you pay the same tuition again.
So before the next bid change, build the handover ledger. Keep the winners with their permission rules. Keep the blockers with their reasons. Keep the watchlist with caps and dates. Then let the new operator improve the system from knowledge, not nostalgia.
Marketplace advertising gets much easier when yesterday’s spend is allowed to teach tomorrow’s budget.