The Business Case for a Customer Portal: How Self-Service Cuts Support Costs and Carries Revenue

A phone or email contact in customer service costs an average of $8.01; the same transaction handled through self-service costs about ten cents — a factor of 80. That 2019 Gartner figure is the most-cited anchor for portal business cases. It is also the reason many of those business cases die internally: promise your CFO a factor of 80, and you are budgeting for savings no portal delivers in year one. This article runs the numbers more conservatively — with figures that still hold up in the second budget meeting.

This piece is written for everyone who has to argue for portal budget internally — as head of sales or head of service facing management and IT, or as a CIO facing the CFO. The business case has two sides: avoided support costs (ticket deflection) and additional revenue (self-service as a sales channel). Both hold up, but at different speeds — and both come with preconditions that rarely appear in vendor presentations. We lay them out openly, including the cases where a portal would be the wrong investment.

The baseline comes from a 2019 Gartner study: live channels such as phone, chat, and email cost an average of $8.01 per contact, self-service channels roughly $0.10. The number is seven years old, but it remains the canonical source — no freely accessible, more recent primary data exists. For your own calculation, the ratio matters more than the absolute value anyway: a live contact ties up staff, a portal lookup does not.

The same Gartner study also contains the inconvenient number: 70 percent of customers attempt self-service at some point, but only 9 percent resolve their issue completely without live contact. The cost lever only exists if the portal actually completes transactions — order status with real ERP data instead of an FAQ page, invoice copies as downloads instead of a contact form. Portal quality determines real deflection, not the mere existence of a login.

Ticket deflection is the share of inquiries the portal resolves instead of your team — the central variable on the cost side. Aggregated benchmark data from 2026 paints a sober picture: in B2B SaaS, the median for realistic deflection through AI-assisted self-service is around 22 percent, best-in-class teams reach 35 to 45 percent, and first-year teams often manage only 10 to 15 percent. For context: these are vendor aggregations, not primary studies — but they sit well below the 30 to 50 percent from vendor marketing, and that is exactly why I consider them the better planning basis.

The second lever is organizational, not technical. According to a Gartner survey from June 2025, 60 percent of service agents fail to actively promote self-service. A portal that nobody on your own team mentions will not hit its deflection targets — which is why the rollout belongs in your service processes and performance targets, not just in the IT project.

And deflection requires integration depth. The Sana Commerce B2B Buyer Report 2025 names missing transparency on stock levels and delivery dates as the top frustration of 40 percent of B2B buyers. That transparency only comes from connecting the portal to your ERP and CRM — how that works architecturally without rebuilding the legacy system is covered in our guide to SSO and ERP integration for customer portals.

Deflection-capable are recurring, read-only, and rule-based transactions: order and delivery status, invoice and document copies, master data changes, standard reorders, ticket status. Not deflection-capable: complaints requiring judgment, consulting, escalations. Those contacts should keep landing with your team — just without the background noise of status inquiries around them.

And measure deflection honestly: a baseline from your ticketing system before launch, categorization of tickets by portal suitability, then a comparison of portal-suitable ticket volumes against actual portal usage. Whoever defines what counts as a deflected contact only after launch produces numbers for wishful thinking — and loses exactly the internal credibility the business case needs in year two.

The cost side carries the break-even; the revenue side makes the case attractive. According to McKinsey's B2B Pulse 2024, e-commerce/self-service and remote sales together account for around 34 percent of B2B revenue — the "rule of thirds." Even more remarkable is the willingness to spend: 39 percent of B2B buyers are prepared to spend more than $500,000 per order through self-service or remote channels; in 2022, that figure was 28 percent.

The preference behind it is consistently measurable. In Gartner's sales survey from June 2025, 61 percent of B2B buyers preferred a buying experience without a sales rep; in the follow-up published in March 2026, it was already 67 percent. The Sana Commerce B2B Buyer Report 2025 — with German buyers in the sample — adds: 73 percent of B2B buyers prefer purchasing online, and 75 percent would consider switching suppliers over a poor digital experience. So the revenue side also has a defensive component: it is not just about cross-sell, but about existing customers who will otherwise order elsewhere.

This expectation is not new: as early as a Microsoft survey published in 2019 (via Statista), 88 percent of customers expected an online self-service offering. What is new is the channel shift behind it — Gartner expects (as of August 2025) that self-service and live chat will overtake phone and email as the most-used service technologies by 2027. Anyone planning a portal in 2026 is no longer building a lead — they are closing a gap. For the internal argument, that helps more than it hurts.

For the business case, this means: reorders that land in your inside sales inbox today are not just cheaper to handle in the portal — they run through the channel a growing share of your customers explicitly prefers. Still, the revenue effect can hardly be quantified credibly in advance, which is why it deliberately does not appear as a euro figure in our sample calculation. It is the cushion above the cost-side break-even, not its precondition.

The following calculation is an example, not a quote — every assumption sits openly next to it so you can replace it with your own numbers. We use a mid-sized B2B company with 2,000 support contacts per month and a fully loaded cost of €10 per live contact (our own assumption, in the ballpark of the 2019 Gartner figure). For deflection, we apply the conservative benchmark values: 12 percent in year one, 22 percent from year two. The investment figures are our published price anchors.

In this calculation, the cost-side break-even lands in year two to three — without a single euro of revenue effect. The sensitivity is worth a look: if deflection stays permanently at the year-one level of 12 percent, avoided contact costs add up to roughly €86,000 over three years — still positive at the lower end of the TCO range, roughly a wash at the upper end. That is the honest reading: a portal that does not get past 12 percent after a year has an execution problem, not a math problem — usually missing ERP depth or a service team that does not champion the channel.

If the numbers come out tight, the full investment does not have to drop at once. A portal MVP — login plus one or two self-service functions with real ERP data — sits between €8,000 and €20,000 at our published anchors and goes live in six to twelve weeks. That lets you validate deflection against real user behavior before commissioning the expansion stage with a role model and deeper integration. From a CFO's perspective, that is the difference between a bet and a staged investment with an exit option.

The math scales down poorly. At 200 contacts per month, even 22 percent deflection frees up just over €5,000 per year — no custom portal can be justified against that. Below roughly 1,000 support contacts per month, or when standard support without ERP depth is enough, we regularly recommend a SaaS solution in the discovery phase: a help center like Zendesk starts at €19 per agent per month according to Zendesk's price list (as of July 2026, billed annually) and covers this case without tying up development budget.

Two more cases without a portal case: if your customers are one-time buyers and there are no recurring transactions — no reorders, no ongoing contracts, no invoice history — the real goal is usually visibility and lead generation; then you do not need a login flow, you need a good B2B website. And if the users would be your own employees — replacing Excel, internal processes — you have taken a wrong turn at the customer portal: that is an internal web app with a different ROI logic.

Take the table above and replace two numbers: your contact volume from the ticketing system and your fully loaded cost rate from controlling. If the cost-side break-even comes in under three years, a discovery is worth it — including an honest make-or-buy check in which we will also tell you if a SaaS solution is enough. How we build customer portals on Next.js, Supabase, and Keycloak, and what they cost, is on our customer portal development page. Or bring your numbers straight to a free intro call — 30 minutes, and you will know whether your case holds.

As of July 2026. Sample calculation with disclosed assumptions, not a quote; all prices net of VAT. Deflection benchmarks come from vendor aggregations, not primary studies.