PDF or web page? What the proposal format actually changes in B2B sales

For twenty years the question about proposal format was a question about looks: nicer template, better typesetting, cover page or no cover page. That question is now obsolete — not because PDFs got ugly, but because what happens to a proposal after sending changed, and so did who reads it.

What actually broke about the PDF

A PDF is not a bad format. It is a format designed to look identical everywhere and never change. Those are excellent properties for a contract. For a sales document they turn out to be expensive in three specific places.

A correction creates a second version of the truth. You spot a typo in the pricing table two hours after sending. You send a corrected file with "apologies, correct version attached". From that moment there are two similarly named files in the client's inbox, and you have no idea which one gets opened in the decision meeting. On an $80k deal that is not a detail.

You know nothing. Opened? Read to the end? Forwarded? Did they stop at pricing or scroll past it? You do not know, so you write the follow-up based on the only information you have — elapsed time. That is where "just checking in" comes from, the most useless sentence in B2B sales. It carries no information because the sender has none.

The phone. The first open often happens on a phone: in transit, between meetings, in the evening. A PDF on a phone means pinching, scrolling sideways, and a pricing table that does not fit the screen. Nobody reads that carefully. At best the client defers it to "later, at my desk" — and later is every proposal's biggest competitor.

Three things a page does that a file cannot

The difference comes down to one sentence: a file is a copy, a page is a place. You can forward a copy, but you cannot update it or learn what happened to it. Everything else follows.

	PDF proposal	Web proposal
Fix after sending	New file, new email	Same link, current content
Know if it was opened	No	Yes, with a timestamp
Know what they cared about	No	Yes — sections and time
On a phone	Pinch, zoom, scroll sideways	A normal page
Video, clickable pricing	No	Yes
Works offline	Yes	No
Contract annex	Native format	Needs an export

Why this started mattering now

Because the buyer increasingly decides without you. Gartner's research puts the share of the purchase journey spent meeting suppliers at 17%, and a typical buying group at six to ten stakeholders. Most of those people were never on your call and will never hear your argument live.

That changes the proposal's job. It stops being a summary for someone who was in the room and becomes the document that has to rebuild the decision context for someone you have never met. And if that is the job, it is worth knowing whether they picked it up at all.

Gartner's more recent survey of 645 B2B buyers (fielded across late 2025) shows two things at once that are easy to confuse: 67% of buyers would prefer to buy without a sales rep, while 69% prefer to validate AI-generated insight with one. That is not a contradiction. It is a precise description of what a proposal has to do now: work on its own when nobody is in the room, and have a human next to it when a doubt shows up.

What knowing changes in practice

The common objection to proposal tracking is that it feels creepy. That would be fair if the point were pressure. The point is far more mundane — it is to stop guessing.

A concrete case. The client opens the proposal on Tuesday and reads for seven minutes. They come back Thursday and spend four minutes on pricing. They open it again Friday morning from a different context. Without that data you write on Monday: "did you get a chance to review the proposal?" With it, you know the deal is live, that price is the topic, and that someone new is reading. So you write about scope options and offer a short call with whoever joined.

That is the whole difference: not whether to follow up, but when and about what.

Where the line is

It is worth drawing it yourself before a client draws it for you. A reasonable rule: a proposal should neither write nor read anything on the recipient's device. No cookies, no localStorage, no browser fingerprinting. Engagement can be computed server-side from data the request already carries, and the reader identifier can be scoped to a single proposal so it cannot be used to follow one person across different deals.

The technical consequence is pleasant: such a proposal needs no consent banner, because it does nothing on the client's device. The ethical consequence matters more: you know your proposal is read like a document, not like an ad-funded page.

When the PDF still wins

Honestly — in three situations, and they are not rare:

1. The proposal is a contract annex. A frozen, unchangeable version is then a feature. You are signing a specific document, not "whatever is currently at that link".
2. Procurement requires a file format. Public tenders and large-company purchasing processes can be inflexible, and there is nothing to argue about.
3. The recipient works offline. Manufacturing, construction, air-gapped institutions — if the proposal has to be read without internet, a file is the only option.

Which is why the sensible setup is not either/or: the page as decision material, an export when formality requires one.

How to test this on your own deals in twenty minutes

You do not need to implement anything to find out whether this problem is yours.

1. Take your last three sent proposals. Count how many went out in more than one version.
2. Count how many follow-ups you wrote without knowing whether the proposal had been opened.
3. Open the most recent proposal on your phone, the way a client would, and read the pricing section.

If the first number is above zero, the second equals the number of proposals, and the third required zooming, you do not have a copywriting problem. You have a format problem.

In short

Proposal format stopped being an aesthetic decision because it stopped being about looks. It is about whether the proposal can change as the conversation changes, whether you know when to come back, and whether it can be read where clients actually open it. The PDF does one thing superbly — it freezes a document — and where that is the point, it should stay.