What 10,000 form submissions taught us about how people actually write to businesses

A while ago I got curious about something. Not curious in the "I should blog about this" way. Curious in the "I can't stop thinking about it" way. When real people fill in a contact form, what do they actually write? Not the marketing version. Not the sanitized case-study version. The real one, with typos and unfinished thoughts and weird openings.

So we took an anonymized slice of 10,000 real submissions from across FormTo customers — stripped of any identifying data, aggregated, run through a bunch of boring counting scripts — and we sat with it for a week. Here's what we found.

A note before I start: none of this includes any personal data or anything that could identify a sender or a recipient. We worked with aggregate patterns only, and we deleted the working set after the analysis.

The average submission is shorter than you think

Here is the distribution of message lengths, measured in words, across 10,000 submissions.

Length	Share
Under 10 words	14%
10–25 words	31%
26–50 words	27%
51–100 words	18%
101–200 words	7%
Over 200 words	3%

The median is 28 words. The average is 47, dragged up by a long tail of people who wrote what are basically personal letters.

Twenty-eight words is half a tweet, roughly. It's enough for "hi, I'm thinking about buying your thing, does it work with Shopify? Also what's the price for 5 users?" That's the actual shape of the average message.

This has a specific implication for how you design the form. Your <textarea> does not need to be 400 pixels tall. It does not need to say "tell us about your project in detail." The people who want to write an essay will write one. Everyone else will write two sentences, and making the textarea feel like an exam question is a small friction that convinces some of them to close the tab.

The opening words are wildly repetitive

I expected a long tail of creative openings. What I got was a bell curve with the same three words at the top.

The most common first words across all 10,000 submissions:

First word	Share
"Hi" / "Hello"	34%
"Hey"	11%
"I'm" / "I"	9%
"Good" (morning / afternoon / day)	7%
"Dear"	4%
"Can" / "Could"	4%
"Is"	3%
A question word (what / how / when / why)	3%
A first name	2%
No greeting, straight into the message	23%

Twenty-three percent of people skip the greeting entirely and open with a question or a statement. This is the most confident group in the dataset, and — anecdotally, from what the senders of those 10,000 messages eventually bought or didn't buy — they're also the group with the highest intent. "Skips the pleasantry, asks the exact thing" correlates with "will pay."

The "Dear" group is older or writing from a professional context. 4% is a small but persistent minority, and they disproportionately come from enterprise emails and government domains.

The things people apologize for (but shouldn't)

We looked for apologies — phrases like "sorry to bother," "I hope this isn't a silly question," "I'm not sure if you're the right person." These appeared in 8.4% of messages. That's roughly one in twelve.

The distribution of what people were apologizing for was striking:

• Asking a question that's "probably already in your docs" — 34% of apologies
• Writing outside business hours — 21%
• Not knowing the right terminology ("I hope I'm using the right words for this") — 18%
• Asking about pricing — 11%
• Being a "small customer" — 9%
• Other — 7%

A third of people who apologized were asking questions they could have answered in your docs, which tells you something about how your docs are organized, not about the user. A fifth were apologizing for writing at 10pm, which tells you something about how hard it is to get rid of the mental model that companies are closed after 5.

The saddest category to me was 9% — people apologizing for being small. "I know we're only a 3-person company, so I'm not sure if this is worth your time." You never want a customer to write that. It's a signal that your marketing is overcorrecting toward enterprise.

What people are actually asking about

We categorized every message into one of eight buckets. This is approximate — multi-topic messages got counted in their primary category.

Category	Share
Pricing questions	22%
"Does your product do X" (feature questions)	19%
Integration questions	15%
Support / bug reports	13%
Demo or call requests	9%
Refunds, billing, account issues	8%
Partnership / business development	5%
"Just saying hi" / fan mail / career / other	9%

A few notes on this.

Pricing is the #1 category, consistently. If your pricing page is hard to find, is vague, or says "contact us for a quote," this number jumps to 30%+. If your pricing page is clear and complete, this number drops to 14%. The contact form is absorbing uncertainty from your pricing page in a ratio you can measure.

Integration questions are chronically under-served by docs. 15% is a lot of messages asking "does this work with Zapier / Notion / Webflow / HubSpot." Most of these can be answered by a single "Integrations" page with a searchable list. The fact that so many still end up in contact forms is a docs UX problem.

"Just saying hi" is 9%. These are the messages I love. Someone reads your blog, likes what you wrote, wants to tell you so. They are not qualified leads. They are better than qualified leads. They will talk about you to other people for years. Reply to them.

The things people leave out

This was the weirdest part of the analysis. We went looking for what wasn't in the messages, and the answer surprised me.

Nobody says their company size unless asked. Of the ~1,000 messages about demos or pricing, only 12% voluntarily mentioned how many people were at the company. The other 88% assumed — correctly — that if it mattered, you'd ask later. The long "tell us about your company" fields on contact forms are extracting information that users don't think is theirs to offer upfront.

People rarely explain their budget. 4% of pricing-question messages included a budget range. The other 96% want to know your number first so they can decide if it fits theirs. Contact forms that ask "what's your budget?" are fighting a norm users have already established.

Deadlines come up only when they're real. 11% of messages mentioned a deadline or timeframe ("we're launching next month," "need this by Q3"). These are high-intent messages. If your form captures the message but not the timeframe, you're losing one of the most useful pieces of signal available for free.

What the top 1% of messages look like

We pulled out the 100 longest, most detailed messages — the ones over 300 words that read like genuine briefs. Here's what they had in common:

• They almost always included a link to the sender's own site
• They usually had a specific question and a specific constraint
• They frequently referenced something the company had written ("I read your post on X")
• They often started with context before the ask ("I'm evaluating three tools, and you're on the shortlist because…")
• They rarely apologized
• They were disproportionately from European senders, especially Nordic and German — culturally, the "long, formal, well-structured" message is much more common there

These are the highest-value submissions in the dataset. If you get one, cancel your next meeting and reply properly. The person who writes 400 words to a contact form has already done more homework than the average sales call.

The boring takeaways

A few things we changed in our own product and process after this analysis, in case any of them are useful:

1. We shortened our textarea placeholder from "Tell us about your project" to "What's on your mind?" — average message length stayed identical, completion rate went up 6%.
2. We added a "link to your site (optional)" field — about 30% of people fill it in, and it makes replying dramatically faster because you can see who they are before you respond.
3. We stopped asking for company size in contact forms and moved it to the demo-request flow where it belongs.
4. We built smarter auto-categorization on the FormTo dashboard so that pricing questions, integration questions, and support can be triaged differently at a glance.

Most of all, reading 10,000 real messages reminded me that the people on the other end of a contact form are shorter, kinder, more hurried, and more apologetic than most marketing assumes. The form is not a transaction. It's a small act of trust. Treating it like a lead capture form is a choice, and often not a good one.

---

This analysis only exists because FormTo customers trust us with their submissions and we take that trust seriously. If you want a form backend that lets you search, categorize, and actually read what people send you — instead of dumping it into an inbox you'll open next Tuesday — start with a free form.

And if you haven't already, read the 30-day spam experiment — the bot data is what this human data looks like when you remove the filtering layer. Night and day.