Supplier A spells it lilac, supplier B violet, supplier C plain purple. To your customer it is one box in the colour filter, but in your catalogue those are three different values. You have two options: spend an afternoon on find-and-replace in Excel, or raise a ticket with the developer and wait three days. And next quarter, when yet another supplier joins, it all starts from scratch.
That is the honest secret behind a lot of product data automation: it is manual work in disguise. There is no system keeping the rules in check, just a person repeating the same action time after time. Every new supplier, every odd spelling, every missing dimension is a fresh little chore. That does not scale, and it is endlessly tedious to boot.
There is a third way. You set the rule once, in plain language, and from then on it runs by itself across your whole catalogue. No code, no queue at IT. In this piece I show how that works, where the limits are, and why automating does not mean handing over control.
What product data automation without code really means
The confusion sits in the word automating. Many people hear "the computer just runs loose". That is precisely what you do not want with your range. Automating without code means something else: you set the rule, the system watches it and applies it, and you stay in charge. From configuring once to running continuously.
The classic building block for that is the decision table. In a decision table you capture business rules in table form: an input condition on the left, a resulting action on the right. In The Power of Decision Tables for Automating Business Rules (2025), Camunda explains exactly why this works for non-technical teams: you can define, update and manage logic without writing a single line of code. So the purchasing or e-commerce staffer sets the rule themselves, rather than waiting on a developer.
// IF input ... THEN valuelila | violet | purple -> Purpleantraciet| DK-GREY | dark grey-> Dark greyecru | offwhite | crème -> Off-white(empty) -> flag for review// 47 spellings in, your 12 colours out
A decision table: on the left what a supplier delivers, on the right what lands in your catalogue.
You read a table like that without a manual. And that is the point: a rule your colleague can read, your colleague can also maintain. Does a supplier turn up next week calling their colour aubergine? One extra row. No ticket, no release, no waiting.
Building an automation: three rules, three minutes
The abstract becomes concrete the moment you build a few. Take a furniture wholesaler who has just received three new supplier files. The brand column is sometimes empty, the weight is in kilos, and the titles are shouting in capitals. Three rules solve that.
These are three of the fifteen ready-made building blocks you use to deploy the 15 ready-made automation building blocks: from default value and find-and-replace to building SEO and the AI assistant. You set them once and they run continuously, per supplier or across your entire catalogue. Importantly: they run on your structure, not the supplier's.
Normalising at scale
Normalising is the dullest and most valuable building block. As long as you have a few hundred items, you can keep spellings in order by hand. At forty suppliers you no longer can. Before long you have forty-seven ways to spell a colour, and just as many notations for a simple dimension.
A rule brings that variety back to your own vocabulary. Millimetres become centimetres, one dimension field becomes height, width and depth, and DK-GREY becomes Dark grey. Not because the supplier delivers it that way, but because that is how your customer filters.
AI as a building block, not a black box
Some fields are not a matter of looking up but of writing. That is where AI enrichment comes in, and it is where many people rightly turn wary. Rightly so, because an AI that rewrites your entire catalogue unprompted is a risk, not a tool.
That is why the AI assistant in SyncRefine is a building block among the others, with the same rule logic around it. It writes descriptions, fills in attributes, suggests categories, translates into 42 languages and even reads attributes from a product photo. What it does not do: take over your writing and tone of voice. It learns your structure, not your voice. And every suggestion is a suggestion: you approve before anything goes live.
That manual work is no detail. According to Parseur's Manual Data Entry Report, staff say they spend more than nine hours a week manually transferring data from emails, PDFs and spreadsheets. AI-assisted enrichment can, according to Crystallize, cut the time to get product content ready from around twenty minutes per SKU to about two. Not because the machine is smarter than you, but because it takes over the mind-numbing part.
Automating without breakages
This is where professional automation parts ways with the reckless kind. A rule that runs blind is a rule that one day sets your entire stock to zero or pushes a price into the shop ten times too high. That is why there is a gate between your rules and your shop.
Quality gates hold anomalies back before they go live: a duplicate SKU, an unusual price or stock jump, a mass deletion. You set them to strict, monitor or automatic. Every push takes a snapshot beforehand, so a rollback is always possible. And you approve. Automating here therefore does not mean less control, but more: anomalies surface before a customer sees them.
- Every odd spelling is a ticket with the developer or an afternoon in Excel
- A wrong price or zero stock is live before anyone notices
- Next quarter your data degrades again, and you start over
- You set the rule once, in plain language, and it runs continuously
- Quality gates hold anomalies back and every push has a snapshot
- Rules watch your catalogue continuously, per supplier or across everything
That continuous character is no luxury. Product data ages constantly: according to Envive, an e-commerce catalogue degrades by roughly twenty to twenty-five percent per year through new products, changing suppliers and shifting specifications. A one-off clean-up does not fix that. A rule that keeps running does.
Automating is not letting the computer loose on your catalogue. It is setting the rules and letting the machine watch them, while you stay in charge.
From source to channel: the same rule, the right field
A clean record is only useful once it reaches your channels the right way. WooCommerce wants different fields from Bol.com, and Google Shopping a different format again from Amazon. That too is a rule you set once and that then fills in exactly the right fields per channel.
SyncRefine connects live with WooCommerce and Magento 2, with Shopify almost ready and Akeneo under construction. Outbound, feeds go to Bol.com in CSV, Amazon in TSV, Google Shopping in XML and Vergelijk.nl in XML. The right format per channel, and only changes are sent. See how SyncRefine takes your product data from source to channel if you want to follow the whole chain.
One raw row, the same rules, and on the other side a record that is right for every channel. Click to see the difference:
{"title": "Oak dining chair","brand": "House brand","colour": "Dark grey","weight": "12,4 kg","shipping_class": "Heavy","height": "48 cm", "width": "92 cm", "depth": "52 cm"// rules applied · approved · ready for WooCommerce}
Click to switch: an empty brand gets the default, DK-GREY becomes Dark grey, and the weight determines the shipping class.
You are not starting from zero
The biggest fear with automating is that you first have to spend months configuring before anything works. That is not necessary. The system reads your existing shop setup, builds a customer profile from it and connects new sources to it non-destructively. Your existing structure is the starting point, not something you tear apart.
With a good network connection the system reads in up to around a hundred thousand products in roughly ten minutes. After that you start small: one source, a few rules, check whether it adds up, expand. You do not have to overhaul your entire catalogue at once to see value.
Take Giga Meubel as an example: a furniture wholesaler with forty suppliers and seventy thousand products live in sync. No invented ROI promise, but concrete proof that rules run at this scale. Want to know how it plays out for your range? Take a look at the frequently asked questions about automations and rules.
What it delivers
Less manual work is the visible gain, but not the only one. Better data sells better: according to Envive, eighty-seven percent of shoppers call detailed product information an important purchase factor, and eighty-three percent leave a site with insufficient information. Around twenty-three percent of returns stem from incorrect product information, and accurate data lowers the return rate by roughly twenty percent.
And speed counts too. Because supplier photos are converted to right-sized WebP on the way in, your pages load faster. Google advises a Largest Contentful Paint within 2.5 seconds, and Core Web Vitals have been a confirmed ranking factor since June 2021. According to the 2025 Web Almanac, only sixty-two percent of mobile pages achieve a good LCP score, so there is ground to be won there.
- A lot of product data automation is manual work in disguise: every odd spelling becomes a ticket or an afternoon in Excel.
- Without code means: you set the rule in plain language, the system watches and applies it, you stay in charge.
- The decision table (if-then) is the core building block; fifteen ready-made building blocks run continuously across your whole catalogue.
- Quality gates and snapshots hold anomalies back before they go live, with your approval as the final step.
- You are not starting from zero: the system reads your existing setup and builds on it. Starting small is fine.
Automating without code is not a promise that the work disappears. It is a promise that the boring, repeated part disappears, and that you keep the rules. You can try it free for 14 days. Start with one source and a handful of rules, and see how much keeps going right on its own.
