DEV Community: Quinn Bennett

ERP, PIM or e-commerce: who should own your product data?

Quinn Bennett — Fri, 06 Nov 2026 16:49:18 +0000

A practical way for B2B companies to split responsibility between systems, so the same product stops existing in four versions.

Ask a B2B manufacturer where the "real" version of a product lives and you often get three answers. The article number and price are in the ERP. The marketing text sits in the e-commerce platform. The latest datasheet is in someone's inbox. Each team is doing its job, but nobody owns the product as a whole.

The result is familiar. A customer portal filters on a dimension that is missing for half the range. A reseller receives an image of the wrong variant. A new market cannot launch because the translation was never approved. None of these are technology failures. They are ownership failures.

The fix starts with a simple question asked field by field: which system is allowed to change this value?

What the ERP should own

The ERP is the system of record for anything that affects money, stock or delivery:

Article numbers and the basic article master
Prices, customer-specific price lists and discounts
Stock levels and availability
Orders, invoices and delivery status

These values change because of business events, not because someone edited a web page. If the web shop is allowed to overwrite them, you will eventually sell something at the wrong price or promise stock you do not have. The ERP should push them out; other systems should read them.

What PIM should own

A PIM (Product Information Management) system owns the sellable, channel-ready description of the product:

Marketing and technical descriptions in every language
Structured attributes per product family (dimensions, materials, ratings)
Images, datasheets, manuals and certificates linked to the right variant
Rules for what each channel needs before a product may be published

The important word is structured. A good PIM model defines, per product family, which attributes are mandatory, which units they use and who approves them. That turns "is the data complete?" from a feeling into a check. If you are evaluating a PIM implementation, ask to see how a single change travels from entry to approval to publication.

What the e-commerce platform should own

The e-commerce or portal platform owns the customer experience:

Navigation, search and category presentation
The cart, checkout and order flow
Customer accounts, self-service and saved lists
Campaigns and content around the products

It should consume product data from PIM and commercial data from the ERP, not become a third place where product facts are maintained by hand.

Draw the line per field, not per system

The hard cases sit on the border. A product name might be created in the ERP but enriched in PIM. A price is owned by the ERP, but a portal may need to show it only to logged-in customers. Write these rules down field by field: source system, who may edit, where it is published. A one-page table with twenty rows will settle more arguments than any architecture diagram.

Make the integrations boring

Ownership only works if data moves reliably between the systems. That means:

One direction per field. A value flows from its owner outwards and is never written back by a consumer.
Clear triggers. Decide whether changes sync in real time, on a schedule or on approval.
Visible errors. A failed sync should land with a named person, not in a log nobody reads.
A plan for upgrades. ERP and platform versions change; the integration needs an owner after go-live, not just a project team that has moved on.

This is where many projects get expensive. Teams build custom ERP integrations from scratch, and then pay again every time the ERP or the platform is upgraded. Pre-built, tested connectors with a named owner after launch are usually cheaper over the life of the solution, even when the first price looks similar.

Three questions to ask before you choose

Whether you are replacing a web shop, adding a customer portal or introducing PIM, these three questions reveal most of the risk:

Who owns each field? Ask the vendor to show the split between ERP, PIM and platform for your own product data, not a generic slide.
What happens after go-live? Who fixes a broken sync on a Friday afternoon, and is that included or billed separately?
What happens at the next upgrade? Are integrations kept compatible with new ERP and platform versions as part of the agreement?

Start small

You do not need to model your whole range on day one. Pick two or three product families, follow them from the ERP to the web, the portal and a reseller, and note every place a value is retyped or overridden. That map is your business case, and it usually shows that the problem was never the number of systems. It was that nobody had decided which one was in charge.

Glanser is a Litium Premium Solution Partner in Jönköping, Sweden. We build B2B e-commerce, customer portals and PIM solutions from a library of pre-built functions and ERP integrations, at a fixed price. More at glanser.com.

Why This AI Tool Works Better Than General Prompts

Quinn Bennett — Tue, 29 Jul 2025 05:39:59 +0000

AI is changing quickly, and many businesses use prompt packs or cheat sheets hoping for quick wins with automation. While these can be helpful, real improvement comes from using AI as part of a larger, well-planned business strategy.

The Problem with Using Prompts Alone

Collecting lots of AI tricks might seem useful, but without a clear plan, you could end up automating problems instead of fixing them. Every business should ask:

Which tasks really benefit from automation? 2.What are the most important goals for the business? 3.Where are you losing valuable time or money?

Relying only on general prompts often leads to scattered results and extra work. To get better results, you need a smart, thoughtful approach.

What Makes This Tool Different

Unlike usual prompt collections, this AI tool gave me something more useful: real clarity. It helped me move beyond just creating random content to building a clear strategy. Here’s what it helped with:
Finding and fixing a slow lead response system, so customers get quicker replies
Automating client onboarding with a small AI assistant to make the process smoother
Creating smart, reusable proposal templates that cover 80% of my work, saving time on repetitive tasks

The Results I Got

Using this tool made a clear difference quickly:
Discovered 6 important AI use cases that fit my business needs
Set up 2 AI solutions in less than a week, saving 6–8 hours every week
Most importantly, I now choose new tools carefully — only using ones that solve real problems instead of chasing every new trend.

A Call to the Community: Let’s Talk AI Strategy

Have you thought about planning how AI fits into your business, instead of just automating for the sake of it? If you use tools or methods that help find the best ways to use AI, I’d love to hear about them.
Please share your experience below. Let’s use AI thoughtfully and with purpose, not on impulse.