You know the drill. A new supplier sends over their product feed — a CSV, an Excel file, maybe an XML dump if you're lucky. Then the real work begins.
Receive the file. Open it. Figure out what the columns mean. Rename them to match your schema. Scan the values for anything weird. Manually clean up colors, sizes, materials. Fix the encoding issues. Convert the units. Test the import. Find errors. Fix those. Test again. Finally, push it into your PIM or shop system.
Two weeks later, the supplier is "onboarded."
That timeline isn't unusual. For most e-commerce teams managing product data, 2–4 weeks per new supplier is the norm. And every monthly update from that supplier still takes a day or two of manual work.
It doesn't have to be this way.
Why It Takes So Long
The onboarding process isn't inherently complex. The individual steps are straightforward — rename a column, translate a value, convert a unit. The problem is that every step is manual, every time.
- No reusability. The column mappings you figured out last month? They're in a spreadsheet tab somewhere. Or in your head. Or in a colleague's head. Next time this supplier sends a file, someone has to reconstruct the same logic.
- Knowledge lives in people. Maria knows that Supplier A calls color
Farbeand uses German values. Tom knows that Supplier C sends weights in grams with no unit label. When Maria or Tom are on vacation, onboarding stalls. - No compound learning. Onboarding your 20th supplier takes just as long as onboarding your 1st. The team hasn't built any infrastructure — just tribal knowledge and scattered notes.
- Error-prone handoffs. When the cleanup involves copy-pasting between spreadsheets, VLOOKUP chains, and manual find-and-replace, mistakes happen. And you often don't catch them until the data is already in your system.
The root cause is simple: the work is repetitive, but nothing captures the logic so it can be reused.
The 15-Minute Onboarding
Here's what supplier onboarding looks like when your mapping logic is saved and reusable. This is a walkthrough of onboarding a brand-new supplier for the first time.
Step 1: Upload the supplier's file (2 minutes)
Drop in the CSV, Excel, or XML file. The system reads the columns and shows you a preview of the data. No reformatting needed — it handles whatever structure the supplier uses.
You select which supplier this file belongs to (or create a new one). That's it for this step.
Step 2: Map columns to your schema (5 minutes)
You see a list of the supplier's column names on the left and your canonical attributes on the right. For common patterns, the system suggests mappings automatically: Product Title likely maps to name, Farbe likely maps to color.
You confirm the suggestions, correct any that are wrong, and manually map any that weren't detected. For a typical feed with 15–25 columns, this takes about five minutes.
Step 3: Set value translations (5 minutes)
Now the system shows you the actual values in each mapped column. For select attributes like color, material, and size, it lists every unique value from the supplier's file alongside your canonical options.
Midnight → black. Eiche massiv → solid oak. XLarge → XL.
You work through the list, setting each translation. For a supplier with 30–50 unique values across their select attributes, this takes about five minutes. Obvious mappings are suggested; ambiguous ones are flagged for your decision.
Step 4: Configure unit conversions (2 minutes)
For measurement attributes — weight, dimensions, volume — you specify what unit the supplier uses and what unit you need. cm → mm. lbs → kg. inches → mm.
If the supplier embeds units in the values (like 40 cm), the system detects that. If they use a separate unit column, you map that too. Quick and mechanical.
Step 5: Review and export (1 minute)
The system shows you the fully normalized output. Every column renamed, every value translated, every unit converted. You scan the preview, spot-check a few rows, and export.
Total time for the first feed from this supplier: roughly 15 minutes.
And here's the important part: the next time this supplier sends a file — next month, next quarter, whenever — you upload it, the saved rules apply automatically, you review the output, and you're done. 2–5 minutes. If the supplier hasn't changed their format (and they usually don't), there's almost nothing to do.
What Makes This Possible
The key concept is what we call a Supplier Adapter — a saved configuration that captures all the mapping logic for a specific supplier.
An adapter stores:
- Column mappings (supplier column → your attribute)
- Value translations (supplier value → your canonical value, per attribute)
- Unit conversions (source unit → target unit, per measurement attribute)
- Any custom rules specific to that supplier's data quirks
When you onboard a supplier, you're really building their adapter. The 15 minutes you spend isn't throwaway cleanup work — it's configuration that persists. Every future file from that supplier flows through the same adapter, producing the same consistent output.
If the supplier changes their format (adds a column, introduces new color values), the adapter flags what's new so you can update the rules. Everything else continues to work untouched.
The Compound Effect
This is where things get interesting. The value of saved adapters isn't just about one supplier — it's about what happens as you scale.
After onboarding 10 suppliers, you have 10 adapters. Each one captures the full mapping logic for that supplier. Monthly updates from all 10 suppliers become a 30-minute task instead of a 20-hour one: upload each file, review each output, export. The adapters do the heavy lifting.
After onboarding 25 suppliers, the math gets even better. Your quarterly catalog refresh — the one that used to block two people for a full week — now takes an afternoon.
And the adapters do more than save time. They also:
- Eliminate knowledge silos. The mapping logic isn't in Maria's head anymore. It's in the system. Anyone on the team can process any supplier's feed.
- Ensure consistency. The same rules run every time. No more "Tom maps it this way, Maria maps it that way" discrepancies.
- Make onboarding a junior task. Processing a monthly supplier update goes from a skilled task requiring deep product data knowledge to a review task that anyone can handle.
Being Honest About the First Feed
We should be clear: the first feed from a new supplier still takes work. Fifteen minutes is real hands-on time where you're making decisions — confirming column mappings, setting value translations, configuring unit rules.
If a supplier has an unusual format, uses deeply non-standard terminology, or sends data with significant quality issues (missing values, inconsistent encoding, mixed languages within fields), the first feed might take 30 minutes or even an hour.
The point isn't that onboarding is effortless. It's that you're teaching the system, not just cleaning a file. The 15 minutes you spend on the first feed eliminates hours of repeated work on every future feed from that supplier. That's a fundamentally different equation than spending 2 hours on cleanup today, and then spending 2 hours on cleanup again next month.
The value is that you never do the same mapping work twice for the same supplier.