You are on the shortlist. Someone you will never meet has your site open in a tab, next to two competitors, and is deciding whether you are what you say you are.
Then they hit the photograph. Four people in a glass-walled office, laughing at a laptop, wearing hard hats indoors.
Stock photography is the fastest way to undo everything the rest of the page just earned. It does not read as neutral. It reads as a claim that turned out to be false, and once a reader catches one, they start checking the others.
Why does stock photography fail so badly in B2B?
Because the person reading is a specialist, and specialists read images the way they read a drawing.
A yard manager knows what a working drydock looks like at 3pm on a Tuesday. A QS knows what a live site looks like when a floor is being poured. They can tell you the hi-vis is the wrong colour for Singapore, that the hard hat has no company sticker, that nobody is wearing a harness where a harness would obviously be required, that the workshop floor is far too clean, and that the shot was almost certainly taken in Eastern Europe. They do not consciously run through that list. They just get a feeling that something is off, and the feeling is correct.
Nielsen Norman Group’s eyetracking work found the same split from the other direction. On one site, users spent 10% more time on the firm’s real staff portraits than on the biographies sitting right next to them, while on another, posed pictures of models were ignored entirely. On a product page, 18% of viewing time went to the images and 82% went to the text, because the text was where the answers were.
That is the whole mechanism. People look at photographs that contain information and skip photographs that contain mood. A stock image contains no information about you, so it gets skipped at best and disbelieved at worst.
The three that are always wrong
The handshake. Two anonymous forearms in suit sleeves, meeting over a table. It has appeared on so many websites that it has stopped meaning partnership and started meaning we could not think of anything.
The harbour panorama. Beautiful, usually shot at golden hour, and completely disconnected from what you do. A superintendent looking for a repair yard does not need to be told what a port looks like. Half the time it is not even the right port.
The model in the hard hat. Clean overalls, folded arms, gentle smile, tablet in hand, standing in front of machinery she is clearly not operating. This is the one that does the most damage, because it is trying hardest.
What actually earns trust
The actual dock. The actual machine. The actual people, in the actual condition they work in.
Specificity is what does the work, not production value. A slightly underexposed photo of your own crawler crane with your own name on the boom outperforms a licensed image of a better crane that belongs to somebody else, because the first one is evidence and the second one is decoration.
This is one of the reasons Towa Marine leads with systems and equipment rather than atmosphere. The buyer arriving on that site is checking whether a specific kind of marine air-conditioning, ventilation or refrigeration job can be handled, and a photograph of a harbour answers none of that. A photograph of the plant does.
Captions matter more than most people expect. An image with a caption naming the vessel type, the system, the site or the scope stops being an illustration and becomes a small piece of proof. An image with no caption is asking to be scrolled past.
The part nobody wants to say
Photography is the hardest line in any project to get funded, and I have stopped pretending otherwise.
It gets cut for reasons that are not stupid. There is no visible before-state, so it is hard to argue for against a line item that produces a page you can click. Getting a photographer onto a live site means safety inductions, client permissions, 2 hours of access and somebody free to walk them round while work carries on. Some clients cannot show the job at all because the end client will not allow it. And the person approving the budget has usually never seen the same page with real photography and with stock, side by side, so the difference is theoretical to them.
That last one is the real problem, and it is fixable. Put the two versions next to each other once and the argument ends in about four seconds.
We quote photography and art direction per shoot rather than folding it into the build, which is on our pricing page with everything else. Partly for honesty, partly because a separate line is easier to defer than to delete, and a deferred shoot still happens.
What to do until then
Something, rather than nothing convincing.
Use what already exists. There are almost certainly usable photos sitting in a site supervisor’s phone, in a handover report, or in a WhatsApp group. They are unglamorous and often better than anything a stock library will sell you.
Shoot the next job properly on a phone. Go out in the 2 hours after sunrise or the 2 hours before sunset, because midday sun on a site gives you glare and black shadow and nothing in between. Back to the sun, phone held level, no filter, no portrait mode, crop hard afterwards. Get closer than feels comfortable. Detail reads as truth.
Photograph the things nobody thinks to photograph. Equipment plates. Test certificates. A drawing on a bench. A finished weld. These are not pretty and they are extremely persuasive to the person you are trying to reach.
And where you genuinely have nothing, use nothing. Typography, white space and a well-written specification carry a page perfectly well. A confident empty space reads as restraint. A generic photograph of somebody else’s business reads as a bluff, and industrial buyers have been calling that bluff for years without ever mentioning it to you.