In most B2B companies, a sales rep spends half their week on things the customer could do themselves — if only they had somewhere to do it. A customer portal is a good idea exactly when it saves that time, and a terrible one when it tries to be an online shop for 12 buyers.
The question "should we build a B2B portal" usually comes up in an SME when the back office can't keep up with the phone anymore. Customers call about order status, call for duplicate invoices, call to check a price they already have in their pricing sheet, and call to ask whether "those screws we've been buying from you for 8 years" are in stock. The instinct is right. The execution often goes hundreds of thousands sideways.
In this piece we look at when a B2B portal actually pays back, what it should and shouldn't pretend to be, and what it costs in practice at a 20–50-person Polish company.
What a B2B portal is (and isn't)
A B2B customer portal is a private zone for your regular contractors, where they log in and see things that concern them: their orders, their invoices, their pricing, their balance. It's not an online shop. It's not a marketplace. It's not an app you'll find in Google.
The distinction is practical. A B2C shop has to convince a stranger to buy. A B2B portal serves someone who has been buying from you for two years and just wants to free up their own back office. That drives completely different priorities: low entry friction (log in, that's it — no registration, no banner carousel), quick re-ordering of past lines, individual price list, real-time status, PDF invoices on click.
If your sales team spends 80% of its time repeating the same answers to the same 30–60 buyers — a portal is for you. If you sell to random visitors who won't return — a portal is wasted, you need a shop or possibly nothing.
A concrete example — an electrical wholesaler
At one of our clients — a 28-person electrical wholesaler serving ~70 regular buyers (installers, smaller contracting firms) — the sales department had three people on full-time desk plus one "person on the phone". Their work in 2024 looked like this:
- An average of 140 phone calls a day across the team, of which ~60% were questions you can answer in Comarch Optima in 10 seconds: order status, stock availability, last invoice, account balance.
- The team was burning an estimated 25 hours a week on "answer the question the customer could check themselves".
- Email orders arrived in 11 formats (Word, Excel, PDF, scanned slip, WhatsApp photo). Each had to be retyped into Optima — index errors ran at ~3% of lines.
We rolled out a B2B portal integrated with Comarch Optima through its API: contractor login by VAT number, per-contractor price list, cart based on previous orders, order status pulled from Optima every 5 minutes, invoice list with PDF download. Project: 4 months, one senior .NET developer plus one half-time integrator. Cost: in the order of 90–120k PLN (roughly 21–28k EUR).
Six months after go-live:
- 62% of order value went through the portal (mainly those who ordered more than once a week).
- Phone calls dropped from 140 to ~70 a day, but call quality went up — the rep was talking about new projects, not invoice numbers.
- Index errors: from ~3% to <0.5% — the customer clicks the product instead of spelling its name down a phone line.
- The "person on the phone" moved into the sales team proper. The investment paid back somewhere around month 11.
Not every company will see these numbers. But if your starting point looks similar — lots of repetitive questions from the same buyers — the direction is probably right.
When a portal does NOT pay off
A portal is an IT project that costs the equivalent of 60–200k PLN. Don't do it if:
- You have fewer than 15 regular contractors. A better email template and a morning report from the ERP will do. Cheaper, faster, works.
- You sell mostly ad hoc. A customer who orders once a year won't create an account, won't remember the password, and won't learn your UI. They'll pick up the phone.
- Your offer changes with every inquiry (e.g. make-to-order manufacturing, where price = project + negotiation). The portal will show a catalogue that can't be catalogued. Trust drops.
- Your ERP data isn't clean. If the same customer sits in Optima under three names and product codes are duplicated — the portal will broadcast that mess to the customer. Clean first, portal later. We wrote about this in the piece on foundations for AI — the rule is the same here.
The worst portal projects we've seen had one thing in common: someone wanted "what the competition has". The competition may have reasons you don't. Count first — how many hours a week does the sales team lose on things the customer could do alone? If it's <10 — let it go. Above 20 — there's a conversation to have.
What the portal should do in v1
The MVP that actually delivers value without ballooning the project into 18 months:
- Login per contractor (one account per company, optionally sub-accounts for staff).
- Your orders — a historical list with one-click cloning ("order the same as last month").
- Order status pulled from the ERP (Comarch Optima, Subiekt GT, enova365 — they all have an API or a database you can connect to).
- Individual pricing per contractor.
- Stock levels (indicative — "available / low / out", not always exact numbers).
- Invoices — list plus PDF download.
- New order — cart and write-back to the ERP.
What to deliberately leave for later: chat with sales rep, customer-side shipping integration, marketplace, AI recommendations, mobile app. Each of these adds another 2–4 months and each could be your next step on its own — but not all at once.
What it actually costs
For a 20–50-person Polish company, built by a normal team (not a Warsaw startup), order-of-magnitude figures in 2026 are:
- Portal MVP (scope from the previous section, one ERP integration): 60–120k PLN (~14–28k EUR), 3–5 months.
- Hosting and maintenance post-launch: 800–2500 PLN/month (depends on traffic and SLA).
- Year-two development (the small things you didn't think of): typically 20–40% of the build cost.
If you're offered anything under 40k PLN — read it carefully. Either the integration is missing (so you get a portal that knows nothing about the ERP — i.e. "a new website with an order form"), or company-level auth is missing, or someone is betting on tacking on change orders mid-build. On the other side — a 400k PLN quote for a 30-person wholesaler should also raise an eyebrow. Your scale doesn't need a solution designed for a 200-branch chain.
Build vs buy
There are off-the-shelf B2B portals on the market (Comarch e-Sklep B2B, e-commerce modules for Subiekt, PrestaShop plugins). For some companies they're enough — especially if you're happy with the standard feature set and have no unusual processes. For others they're a trap: every non-standard requirement needs a workaround, and those workarounds aren't something you'll build in-house.
Practical rule: if your sales process is "standard" — buy the boxed product. If you have 2–3 things nobody else has (a specific discount system, your own stock reservation logic, an unusual transport integration) — a custom portal comes out cheaper over three years, even though it looks more expensive on day one.
What next
A B2B portal isn't a "technology" project, it's a business one. Whether to do it depends on a handful of numbers worth counting calmly: how many regular contractors you have, how much time per day the team loses on repeat questions, how many orders get lost in inboxes. If those numbers aren't obvious yet — that's a good moment to spend half an hour on it before spending the first zloty.
A B2B portal isn't meant to be a pretty shop. It's meant to be the tool that stops your sales rep from picking up the tenth phone call asking "when will my order be ready".