The San Francisco Check I Run Last, and the Only One That Stung

83.2% of the San Francisco businesses I sampled hold at least one Google review, against 88% across everything I’ve counted. I ran my whole audit on this city and only one check bothered me.

The San Francisco Check I Run Last, and the Only One That Stung

Every job you finish in San Francisco ends in the same awkward minute, and nobody has ever known what to do with it.

The work is finished. You’re out by the van with your phone in one hand, thumbing an invoice into an app, and she’s standing in the doorway with her arms folded, pleased, waiting for a decent moment to say thanks and get back to her evening.

So you tell her to give you a ring if anything plays up. She says she will. You drive off.

That minute goes by a few hundred times a year in every trade in the city. It’s the most valuable sixty seconds in your working week, and nearly everyone hands it back.

Keep that doorway in your head. My audit runs in a fixed order, ten checks and a few extras, and the doorway is where the whole thing ends up.

The order, and the first thing I look at

A few weeks back I put San Francisco through the same run I do for a paying client. One line per business, one mark per check, and the checks go in the order I trust rather than the order that looks tidy.

Before I read a single business, I look at the overall score, because the shape of a city tells me what kind of week I’m in for. San Francisco settled at 72 out of 100, 14 of 24 among the cities I ranked.

Middle of the pack. For a city that exports software to the rest of the planet, that was already more interesting than a win would have been.

Then I go back to the start and work through them properly.

First check: is there a website at all

Nothing clever about the first one. Either the business lists a site or it doesn’t, and in San Francisco 82.7% of my sample lists one.

The blanks are not always what you’d guess. Plenty of them are shops doing perfectly well on foot traffic and a phone number, and a missing website stopped meaning a struggling business a long time ago.

Have you ever counted how many of your own competitors don’t have one? It’s usually more than owners expect, and it matters less than they expect too.

Second: the phone checks, which I skim

Two separate checks live here, and owners mix them up every single time.

One of them is a shape question and the other is a waiting question, and a page can sail through the first while a customer stands on the corner of Geary watching a white screen do nothing.

Both of them come back healthy in this city, so both got a skim and nothing more, which is not something that happens with most places.

Was your own site built for the fit or for the wait? Hardly anyone I ask can tell me.

Third: the code, where I lose an afternoon

On my screen this part is a wall of yes and no, one narrow line per check, and it fills in slowly because a machine has to open every page and read the parts a customer never sees.

No owner has ever rung me up angry about a tag. The code checks here are fine, the same way a well kept van is fine, and there was nothing to complain about by the end of them.

That’s worth saying out loud, because a lot of what gets sold as urgent lives in those checks.

Fourth: the writing

Now it gets more human. Does the business publish anything, ever, that wasn’t written by the person who built the site?

39.7% of the San Francisco sites I opened publish something, and that one made me sit up, because in most cities it’s a graveyard of posts that stopped in 2019.

Someone up here is still writing. Would a customer find anything on your site written this year, in your voice, about a job you actually did?

Last: the listing, and the doorway

I do the review checks last, always, and I’ve never worked out whether that’s good discipline or plain avoidance.

The review check comes back as a pair of figures side by side. How many people have written anything, and what they gave the place out of five.

A business with none shows a blank where the count should be, and blank means a name that turns up in a list of local options carrying nothing underneath it to argue its case.

83.2% of the San Francisco businesses in my sample carry at least one Google review. I read that, shut the laptop, and went and made a coffee.

Because I had it backwards on the drive in. The country as a whole sits at 88%, and I’d assumed a city where people rate their sandwiches would be sitting comfortably above the country, not looking up at it.

The check that costs a business nothing is the one where this city is behind the rest of the country. Every other check I ran that week landed somewhere between fine and good.

Why I think it goes this way here

None of what follows is provable. An audit counts things, it doesn’t explain them, and anyone telling you otherwise is selling something.

But I have three guesses, and every one of them ends up back in that doorway.

  • The ask never gets made out loud. It gets put on a card, or in a footer, or in an email that arrives four days later when she’s stopped thinking about you. Heating companies are the exception I keep pointing at, because they ask on the way out while the house is warming up again.
  • Everyone assumes the platform handles it. Plenty of trades here take bookings through an app, and an app that collects its own ratings feels like the job is done, right up until a stranger searches your name on a map instead.
  • People move. Staff move, customers move, and a business that leans on a good year of word of mouth can find the word of mouth has quietly relocated to Austin.

None of those is a website problem, which is awkward for me, because websites are what I sell.

When someone here hands me a site to rebuild, the first thing I ask for is the review link, and it goes on the thank you page before I’ve picked a single color.

What actually changes if you do it

Ask the next three customers in that minute at the doorway. Out loud, while she’s still pleased with you, phone in your hand so you can send the link before you pull away.

Not a campaign, not a printed card, not a reminder email with a subject line you labored over. Three asks, this week, in person.

If the words are what’s stopping you, I’ve written down how I ask, and saying it takes about ten seconds.

A few of you shouldn’t ask at all. A psychologist told me she wouldn’t put that on a client, and I agreed with her before she’d finished.

My promise is deliberately small. By the end of the month your name will carry a count and a rating where today it carries nothing.

Nothing else changes. Your phone won’t ring off the hook on Friday and no van will need buying.

But the next person comparing three businesses she’s never heard of, on a phone she doesn’t much want to be holding, will have one fewer reason to scroll straight past you. Over a year, how many of those do you think you’ve lost without ever hearing about it?