SaaS Wins Every Check You Can Type and Loses the Ones You Build

A founder I met once at a software conference, for about four minutes, texted me on a Sunday night.

“Three quotes back for the marketing site. Cheapest is a third of the others and they swear it’s the same build. Is it?”

I typed four answers to that and deleted all of them.

So this is the long one, and it’s for you if a quote like that is sitting open in a tab this week.

I’ve answered a version of that text a dozen times, and my answer has never once turned on the price.

Yes, they can build you a website. So can the other two, and so did every competitor you’ve ever lost a deal to, which is why software already finishes first of my 47 industries for owning one.

96.7% of the software companies I sampled have a website, and across small businesses generally it drops to 83%.

Nobody ever won a deal by owning a website. So the quote isn’t really buying you a site, it’s buying you the distance between yours and the one at the company you keep losing to.

I went looking for that distance across 757 audited sites and 565 speed tests run on a phone.

It isn’t budget and it isn’t taste. It comes down to whether the fix is something a person types, or something a person has to build.

The half that gets typed

Every quote you’re holding covers this half, cheap or expensive, because the software drops most of it in before anyone touches the page.

The best sentence on most software homepages is one no visitor ever reads on the page itself. Meta descriptions reach my screen as a stack of one line pitches, one per site, and I can usually tell which companies employ a writer inside the first four.

Software fills that slot more often than the field does: 89.7% have one against 77.9% of every site I audited.

Open Graph tags came back in my audit as a handful of short lines, an image, a title, a line of text.

Leave those rows blank and your pricing page turns up in a chat window looking like a link someone pasted there by mistake.

84.8% of SaaS sites fill them in against 75.7% of everything I audited.

Security told me the same story twice. 97.0% serve over HTTPS against 94% of audited sites, and 94.2% push insecure traffic across to the secure copy against 88%.

Not one of those is more than a line of text in a file. Someone types it once, ships it, and the check stays green forever.

Your trade is very, very good at typing lines into files. Ever noticed that the checks your team aces are the ones one person can finish in an afternoon?

The same shape turns up away from the website entirely, where software blogs and tracks harder than any industry I counted and then leaves its Google profile empty.

The half you have to build

Now the ones nobody types their way out of, and the mood changes fast.

I open a lot of software homepages on my own phone, and this is the half where I stop being impressed.

Speed goes first, and it flattens the whole story. Put the SaaS average of 58 out of 100 for mobile speed beside the 58 out of 100 every audited site averages and there’s nothing left to brag about, which parks software in the middle of the field.

Schema markup is the check where my audit either finds a labelled block of code in the page or finds nothing at all. The sites with nothing are asking a search engine to work out what they sell by reading the sentences, the way a person would.

71.7% of SaaS sites carry it against 70.5% of the sites in my audit, which leaves software level with the field rather than ahead of it.

An industry that runs on product screenshots, and I still open homepages where not one of those images has a word of description attached to it.

See what happened? The second a check needs a change to the way the site is built, first place quietly turns into average.

Why the line falls there

My theory is dull, and I’ve watched it play out on enough projects to trust it.

Your marketing site is an orphan.

The app has owners, a review process and someone on call at three in the morning. The marketing site is a thing a contractor built before your Series A, and it belongs to whoever happens to be least busy this month.

Someone has to own the upkeep, and the list of what upkeep actually covers runs longer than most founders expect.

41.6% of these companies run WordPress, more than any other platform in the industry, with Webflow on 11.8%, Shopify on 11.1%, HubSpot on 5.0% and Framer on 3.4%.

I’ve watched marketers log into all of those and fix a description themselves, with no ticket, no standup and no engineer interrupted.

Speed doesn’t work that way. Neither does the way your images get handled, or the way your pages get labelled for a search engine.

Those want someone who understands how the site is put together, and that person is busy shipping the thing your customers actually pay for.

So the typed half stays green while the built half quietly rots, year after year, and nobody in the building ever gets a notification about it.

Counting goes the same way. Plenty of companies have analytics running and nobody left in the office who can open the account.

When did an engineer last open the repository your marketing site lives in?

What the cheap quote really costs

Does a slow marketing page lose you a deal on its own? No.

But your homepage is the first thing your product ever does in front of a buyer, and it’s the only performance she can judge before she trusts you with anything.

A page that snaps into place on a phone in a parking lot says the company behind it knows what it’s doing.

A page that hangs, then shuffles itself around while she’s trying to tap your pricing link, plants a small doubt she’ll never mention to you or to anyone else.

I can’t put a number on that, and I won’t pretend my data can either.

Rebuilding marketing sites for software companies is my actual job, so read that paragraph with the bias in mind, and what I charge sits on one page if you want it beside the other three quotes.

If it helps to look at finished work instead of averages, I keep a shelf of SaaS site examples here.

Before you reply to any of them

One test first, and it costs you a minute. I run it on my own site twice a year, and it has embarrassed me at least once.

Take your phone off the office wifi and open your marketing homepage on cellular data, somewhere with two bars, the way a buyer sitting in an airport would. Count the seconds out loud, then open your own app on the same connection and count again.

Is the page selling the product slower than the product?

Now put the same question to all three of the people who quoted you, and watch which one has an answer ready: once the typing is done, who owns the parts that have to be built?

Read the three replies next to each other before you sign anything.

So which of them quoted you for a website, and which of them quoted you for the typing?