Quick answer
Name who uses it, whether they log in, and whether a stranger has to trust it. A website is a page a stranger reads and one action they take. A product is a loop a signed-in user returns to, with data that belongs to them. An internal tool is a job your team already does — ugly is fine if it works on Tuesday. We pick the smallest of those three that is still honest. The expensive mistake is building the other two because the brief said “app.”
Key takeaways
- The first question is the job, not the logo. Framer, Lovable, and a custom build all fail if you picked the wrong noun.
- A website sells or explains. Marketing publishes it. Nobody logs in.
- A product has accounts, their data, and a reason to come back. That is a different stack and a different quote.
- An internal tool has three users who already know the process. Ship the job. Do not brand it.
- Most companies need a site and a product. Glue them. Do not stuff one inside the other.
At a glance
| Website | Product | Internal tool | |
|---|---|---|---|
| Who it is for | A stranger, once | A user who comes back | People on your payroll |
| The job | Believe you, then act once | Complete a loop they own | Finish the weekly job |
| Login | No — a form is not an account | Yes, with their data behind it | Often, against your own SSO |
| Who changes it | Marketing, weekly | Product, as usage teaches you | Ops, when the process moves |
| Wrong reason to pick it | "We need to look like a platform" | "The investor asked for an app" | "Let's productise the spreadsheet" |
| Our default | Marketing site in Framer | MVP, then SaaS if it earns | Replit, n8n, or the spreadsheet — on purpose |
If you cannot say which column you are in, you are not choosing a studio. You are buying a build that will answer the wrong brief. We have this conversation before we open a file.
What each one actually is
A website is a designed argument. Pages, a CMS, one action — book, buy, apply, join the list. The visitor does not have an account. The team that writes the copy should be able to publish without us. That is marketing sites. For most of those we default to Framer development. If the incumbent is WordPress or Webflow, read those compares first — the noun is still “website.”
A product is a loop. Sign in, do the job, see your own data, come back. Auth, permissions, a model you own, usually billing. An MVP is the smallest version of that loop that is still real. Custom SaaS is the same job once a template's assumptions break. A landing page with a fake dashboard behind a button is not a product. It is a website that lied.
An internal tool is a process you already run. Three people, a Slack channel, a sheet that has become the company. The user already knows the job. Taste is optional. Reliability on Tuesday is not. That might be a Replit app, an n8n flow, or a cleaned-up spreadsheet. If the only customers are on the payroll, you do not have a SaaS. You have ops.
A workflow that moves data between tools is not a product and not a website. It is workflow automation. We build those in n8n. Do not commission a dashboard because you were embarrassed to say “the bot does it.”
The mix-ups that waste a quarter
Five marketing pages and a “portal” that is a CMS list. You needed a website. The login was theatre.
Accounts, roles, and a billing state on a canvas. The canvas will take the money. It will not be a product. That ceiling is when Framer is not enough.
You want tenants, plans, and a marketing site for a tool only ops will use. Build the internal tool. Skip the pricing page.
The product works. Nobody can explain it. You did not need more features. You needed the website that sells the one you have.
The tell is who you are afraid of. If you are afraid of a stranger bouncing, you need a website. If you are afraid of a user losing their work, you need a product. If you are afraid of someone on your team doing the job by hand again on Monday, you need a tool. Quote the fear, not the moodboard.
How we choose on a call
- 01Name the human and the return visit
If they never come back, it is a website. If they come back to their own data, it is a product. If they are on the payroll, it is a tool.
- 02Ask who publishes next month
A marketer who will not open a repo needs a site with a CMS. A founder who will live in the build can take an MVP. Ops will tolerate a Replit they were walked through.
- 03Ask what happens if it is ugly
If ugly kills the deal, you are selling. That is a website, or a product a stranger will judge. If ugly is fine, do not buy a brand system for an internal job.
- 04Refuse the hybrid with no pain
“Site plus app plus AI” is three briefs. We will cut to the one that is late, or we will sequence them. We will not pretend they are one Framer file.
“If you cannot say whether a stranger, a customer, or a colleague has to use this on Tuesday, you are not choosing a stack. You are choosing a noun.”
Chris NutbeenNuttifox
When you need two
Most companies that sell software need a website and a product. The site is the argument. The product is the loop. Keep them apart. Marketing publishes the site. Engineering owns the app. A shared design system is enough glue. Butter is the commerce version of that split: the store stays a store, the designed surface is Framer.
Do the website first if nobody can explain you. Do the product first if the only people who matter already believe you and need to click the loop. Do the internal tool first if the “product” is still a weekly ritual in a sheet. We publish a $2,000 a week full-time rate on the contact page. A wrong noun is the line item we try to delete on the call.
If you cannot name the user, the return visit, or the Tuesday job, do not hire us yet. Write that down. The first week of an MVP is that sentence. The first week of a site is the one action. The first week of a tool is the process. No file until one of those exists.
FAQ
Do we need a website or an app?
If a stranger must believe you and act once, you need a website. If a signed-in user must come back to their own data, you need a product. “App” is not a diagnosis.Can Framer be the product?
Framer is a marketing canvas with an escape hatch. A calculator or a form can live there. Accounts, roles, and billing cannot — not as the product. That is when Framer is not enough.When is an internal tool allowed to look unfinished?
When the only users already know the job and you would rather have the Tuesday back. The moment a customer or a candidate will see it, you are in website or product territory.Should we productise our internal tool?
Only if someone who is not you will pay, and you can name the loop. Three teams on a shared sheet is not a market. It is a process. Automate it or tidy it. Do not invent a pricing page.What do you actually build for each?
Websites: usually Framer. Products: an MVP, then a custom build if it earns. Internal tools: Replit, n8n, or the sheet. We will say if you are in the wrong column.Can we start with one and graduate?
Yes. A site can stay a site while the product starts. An internal tool can become a product when a stranger will pay. Graduation is a new brief, not a toggle in the same file.Have something to build? Let’s ship it.
Book a call