Decision guide

Website vs client portal vs SaaS

Joel Gascoigne stopped coding Buffer and put up a two-page site, then a pricing page. People paid before the product existed. He has also written about an earlier product he built without that test, and about a Buffer for Business idea his own customers would not pay for. The site was the experiment. The product was the thing he built after the experiment worked.

Written by SiteMivo. Drafted with research assistance and checked against the sources below.

Four ways this purchase goes wrong

A login nobody opens. Steve Smith, who builds client portals, writes that clients do not sign in for a one-way delivery. Nielsen Norman Group’s older research says a login wall before any value stops people. Phelicia Gomez’s Phab Phit site shipped in 2026 without a login, a portal, or booking, on purpose.
A portal for a brochure. If the visitor’s job is to understand the offer and write to you, the signed-in app is a second product you now have to run.
A SaaS build before a payer. Emanuele Di Pietro spent months on DP’s Templates, got no paying users, and Supabase shut the project for inactivity. A developer writing as kennedy reported six months, twelve signups, and zero payers.
An internal tool sold as a product. Consoles your own team uses are not a SaaS until a stranger pays to use them.

Sources: Gascoigne, 2011, Gascoigne, 2014, Smith, Nielsen Norman Group, Di Pietro.

The test is who comes back next week

Question	Website	Portal	SaaS
Who signs in	Usually nobody. They read and enquire.	People you already work with, and they come back.	Accounts that pay, each with their own data.
What ‘done’ means	The right person takes the next step.	A recurring job leaves email.	A stranger finishes the job without you.
First build	Pages that explain the offer.	One workflow, not every process.	Sign-up, the job, a charge.
After handover	Code, domain, content, deploy.	Those, plus roles and their records.	The product, the tenant data, and billing.