Why this signal matters for Google
A product photo you took yourself is proof first and a picture second. Two pages review the same air fryer. The first shows the manufacturer's image on a white background, the one you can find on two hundred comparison sites. The second shows the fryer on a kitchen counter, basket pulled out, a grease mark on the handle, the box still on the floor. Nobody needs to know anything about photography to tell which page actually tested the thing. That is the whole territory of the Product photo signal: not the pretty picture, the evidence that you had it in your hands. So this guide answers the question of how to take product photos differently from the studio tutorials: we want the photo that proves, not the photo that sells. It extends the logic of the original photo, an image that exists only at your place. And readers go looking for it first: in the Baymard Institute usability tests (2020), 56% of users start by exploring the photos when they land on a product page.
Google asks the question in plain words: did the author use the product? Its quality evaluators work from a rulebook, the Quality Rater Guidelines. Section 3.4 defines the first letter of E-E-A-T, experience, in very concrete terms: has the person writing used the product, visited the place, lived the situation? Section 7.3 adds that first-hand experience alone can make a page High quality, no credential required: someone describing what a product did in their home counts as much as a specialist, as long as the proof is there. The full document is online. A photo of the product in your hand, with your floor and your light behind it, answers that question in a second, where 800 words of text leave room for doubt. On the affiliate sites I audit, it is the first test I run: I look for a single photo that did not come from the manufacturer, and often I cannot find one.
What makes a photo believable is exactly what a studio removes. A white background erases context, and context is what proves. Keep the surroundings: the table, the power outlet, the hand holding the object. Shoot several angles, including the ones the manufacturer never shows: the underside, the inside of the compartment, the label with the serial number. Show the flaws: the paint chipping after three weeks, the seal that yellowed. A photo that shows a defect is more believable than a perfect one, because no brand would ever have supplied it. Add time: the same pan photographed on purchase day and six months later tells what no spec sheet can, which is the principle behind the logbook. And keep the EXIF metadata, the information the camera writes into the file (date, phone model, sometimes location): it confirms the photo came from you and not from a stock library. When the product touches health or money, a supplement, a blood pressure monitor, a safe, the page falls under YMYL (Your Money or Your Life), the category where a mistake costs the reader and where Google demands the most proof. A photo of real use weighs even more there.
Answer engines sort lived tests from copied spec sheets, and your photos are part of the sorting. Ask Perplexity whether that air fryer is worth the money: the answer cites pages where someone obviously took the appliance out of its box, because those are the only pages holding information the model does not already have. AI systems read the alt text and the caption of your images; Google's visual search recognizes a product in a photo and spots an image it has already seen elsewhere. A reused manufacturer shot tells these systems there is nothing new here. A photo taken in your garage, captioned after two months of use, says the opposite. And an AI-generated product photo does the exact reverse of what you need: it looks like a packshot, carries no camera and no date in the file, and proves only that you can write a prompt, not that you ever held the object.
The NeuroEEAT scoring grid (0-10)
Score your product photos on what they prove, not on how good they look: could a stranger swear you had the object in your hands?
| Score | Description | Concrete indicators |
|---|---|---|
| 0-2 | Manufacturer photos | Every image comes from the brand or a stock library. White background, no hand, no context. The page could have been written without ever seeing the product, and that is what a reader concludes. |
| 3-4 | One home photo, wasted | One or two photos you took, but cropped to look like a packshot: neutral background, a single angle, metadata stripped on export. Nothing sets it apart from an image found online. |
| 5-6 | The product in context | Several photos at your place, the object in use, at least one angle the manufacturer does not show. Captions stay vague and the photos carry no date. |
| 7-8 | Complete proof | Visible surroundings, multiple angles, flaws shown, captions that say when and where. Files keep their EXIF data and the author signing the text is the one holding the object. |
| 9-10 | Proof over time | Unboxing photos followed by photos after months of use, compared side by side. Receipt or serial number shown where useful, ImageObject markup with the author, and a published rule: no photo without a real purchase or loan. |
How to audit it on your site
Take your most-read product page and put these questions to it, in this order.
- 01
Search every image in Google Lens
Run a reverse image search on your photos. If Google finds the same image on thirty other sites, it came from the manufacturer and proves nothing. A genuine photo of use has no duplicate.
- 02
Count the photos that show your surroundings
A hand, a floor, a window, another object of yours in the frame. Zero photos of that kind on a page that claims to test a product is the problem to fix before anything else.
- 03
Find the angle the manufacturer never shows
Underside, inside, label, cable, opened packaging. If every one of your photos repeats the angles of the official listing, an attentive reader concludes you copied the listing.
- 04
Look for a flaw or a trace of use
Scratch, dust, wear, a part that does not sit right. A page showing only spotless new items reads like an advertisement. One photographed defect changes how the whole page is read.
- 05
Open the metadata of one file
Download one of your images and look at its EXIF information (right click, properties, or an online tool). If the date and the camera vanished on export, switch to a workflow that keeps them. If they name an image-generation tool, the photo no longer protects you.
How to activate or improve it
You need neither a studio nor a photographer. You need the product, your phone and a little method.
Take the product out of the box in front of the camera
- Photograph the unboxing: box closed, box open, contents spread on your table. It is the simplest proof and the hardest to fake.
- Shoot at least five angles, including the underside and the label, using window light rather than flash.
- Keep the setting: your kitchen, your workshop, your desk. Do not chase the white background.
- Write a caption that says when and where: photo taken March 12 in my garage, after two weeks of use.
Show what the manufacturer hides
- Photograph the product in use, hand included: the button pressed, the drawer open, the result obtained.
- Document every flaw you find with a close-up and one sentence saying whether it matters.
- Export your images with the EXIF metadata intact; turn off the strip-information option in your compression tool.
- Retake the same photo, same angle, after one, three and six months of use, and show them side by side.
- Put the photo of real use first, ahead of any official manufacturer image.
Make the proof machine-readable
- Add ImageObject structured data to each image (JSON-LD, code the reader never sees) with the author as a Person, the creation date and a license.
- Publish your testing rule on a dedicated page: products bought or loaned, minimum trial period, what happens to the product afterwards.
- Show the receipt or the serial number, partly masked, when the product's value makes proof of purchase useful.
- Keep the untouched original files in a dated archive: that is your defence if a photo's authenticity is ever challenged.
- Review your ten most-read product pages every quarter: each must keep at least three current photos of real use.
5 common mistakes to avoid
Sector specifics
Supplements and measuring devices
A blood pressure monitor or a supplement touches the reader's health, and evaluators look for more proof there than anywhere else. The photo of the box with the batch number, of the screen showing a real reading, of the product next to a blurred prescription: that is what separates a lived test from a sponsored article. Never show a reading or a result you did not obtain yourself.
Expensive equipment and valuables
On pages comparing safes, counterfeit detectors or costly professional gear, a photo of the product in hand with a partly masked receipt answers the question every reader is asking: did this person really buy it, or are they earning a commission on the link?
A store that sells the product
You hold the stock, so there is no excuse for reusing the manufacturer's photo. Photograph your own unit, the tape measure beside it, the product worn or installed at a customer's home with their permission. A listing with your own photos of use stands out from the hundred other listings selling the same reference.
The product installed at the customer's
A heat pump installer or a kitchen fitter shows the unit in place, with the customer's wall, floor and light. It is a product photo, a proof of work and a local proof all at once. Ask permission before publishing, and keep the raw photo with its date.
Signals to activate in parallel
The product photo proves you had it in your hands. These signals say what you did with it:
FAQ
How to take product photos on your phone?
Take the product out of its box onto your table, near a window, and shoot it with your phone from at least five angles: front, sides, underside, detail, product in hand. Keep the setting, do not chase a white background. Under each photo, write when and where it was taken. In ten minutes you have images no other site owns, and that is what counts for trust.
How do I take simple product photos?
Window light, a steady phone, a clean product, one clear angle per shot. That is all the technique you need. Then resist the urge to polish everything: a result that is too clean looks like a bought image. Simple does not mean careless. It means the reader understands the object and believes the person showing it.
Can AI do product photography?
As illustration, perhaps. As proof of experience, no. A generated image has no camera, no date and no real flaw, and it shows a product you never held. If you use one, say so in the caption. For the Product photo signal it counts for zero, and it can undo the trust your real photos had built.
How to take product photos for a website?
Put the photo of real use first, ahead of any manufacturer image, and give it a caption that says when and where. Keep several angles the official listing does not show, keep the EXIF data in the file, and write alt text that describes what is in the frame. That is what a reader, an evaluator and an AI system can all check.
How to photoshoot a product at home?
Home is the point, not the limitation. Your counter, your desk and your garage are the proof that you own the thing. Photograph the unboxing, the product in use with your hand in the frame, the underside and the label, then the same shots again after a few months. A studio would erase every one of those clues.
Case study
Before · a review page built on manufacturer photos
After · the same page with photos of real use
A scenario built to show what changes on the page. The values do not come from a real client and promise no ranking outcome · measure with your own pages.