Your Job Photos Prove You Worked in Manhattan Beach — Every File Google Actually Sees Says Nothing At All
A Palos Verdes landscaping crew finishes a hillside retaining wall job in March. The foreman takes eleven photos on his iPhone — before, during, after, a drone shot of the finished terrace. Three of those photos end up on the company's "Projects" page. Two get posted to Google Business Profile with the caption "Recent work!" The rest sit in a camera roll, doing nothing.
Here's the part the owner never checks: not one of those photos, anywhere they live online, tells a machine they were taken in Palos Verdes. Not Google. Not Bing. Not the AI models now summarizing "best landscaping company near Palos Verdes Estates" for people who never click a blue link. The company did the job. The proof exists. It's just invisible.
Your Camera Knew Where You Were. Your Website Never Found Out.
Every modern smartphone photo is born with GPS coordinates baked into its EXIF metadata — the same hidden data layer that records the camera model, the timestamp, and the exact latitude and longitude of the shot. In theory, that's a gift for a service business trying to prove it works across a wide territory. A plumber who did four jobs in Redondo Beach last month has four pieces of geographic evidence sitting in his phone right now.
In practice, that data dies before it ever reaches a search engine — and it dies in three specific, predictable places.
Strip One: The Phone Itself
The moment that photo leaves the camera roll — texted to the office, AirDropped to a laptop, uploaded through a browser share sheet — iOS and most messaging apps strip the location metadata by default. Apple did this for consumer privacy reasons, not malice, but the effect is the same: by the time the file reaches a marketing folder, the GPS tag is frequently already gone.
Strip Two: The CMS
If the geodata survives the transfer, it almost never survives the website. WordPress optimization plugins — Smush, ShortPixel, Imagify, the tools every site owner installs to fix a slow PageSpeed score — compress images by rewriting the file, and rewriting the file wipes the metadata clean by default. The business fixed one problem (page speed) and quietly created another (geographic evidence) without anyone flagging the tradeoff.
Strip Three: Google Business Profile's Own Upload Pipeline
Even the photos uploaded directly to GBP get flattened. Google's own upload process strips EXIF data from consumer photo uploads across most of its products, including Business Profile. So the two photos captioned "Recent work!" that made it to the map listing arrive with zero location signal attached — just a filename like IMG_4021.jpg and a caption that says nothing about where the work happened.
By the time a photo of a real Palos Verdes job reaches the public internet, it has been geographically scrubbed three separate times.
Google Isn't Guessing Your Service Area — It's Reading Whatever's Left
To be precise: Google has said plainly that raw EXIF GPS data isn't a primary ranking signal for Search. That's true, and it's worth being honest about. But that's not really the failure worth diagnosing here. The failure is what EXIF data represents — a business's only unstructured proof of where it actually operates — and what happens when nothing replaces it.
Local ranking systems and the newer AI-search layer sitting on top of them (the systems producing AI Overviews and the answers ChatGPT gives when someone asks "who does landscaping in Palos Verdes") build their confidence in a business's service area from corroborating signals: the address on the Business Profile, service-area settings, citations, review text that mentions specific cities, and — increasingly, as these systems get more multimodal — the text surrounding images: alt attributes, file names, captions, and structured project data.
Strip the EXIF data and do nothing else, and you've lost one weak signal. Strip the EXIF data and leave the alt text blank, the filename as IMG_4021.jpg, and the caption as "Recent work!" — which is what happens on the overwhelming majority of trades websites — and you've lost the entire chain of evidence. The photo exists. It proves nothing. A model trying to answer "does this landscaping company actually work in Palos Verdes, or just Torrance?" has nothing in that image to synthesize an answer from.
This is exactly the gap that shows up when a general contractor or landscaping company wonders why they rank confidently in their home city and disappear the moment someone searches from three towns over. It's not always a service-area setting or a missing citation — sometimes it's forty beautiful, completely anonymous project photos.
What a Photo Architecture That Actually Proves Location Looks Like
Fixing this isn't about fighting Apple's privacy defaults or begging Google to stop stripping EXIF on upload. It's about accepting that the metadata layer is gone by design and rebuilding the geographic proof in the layer that survives: the text around the image.
Filenames carry the city and the job. retaining-wall-palos-verdes-estates.jpg tells every downstream system something IMG_4021.jpg never will. It costs nothing and takes ten seconds per file.
Alt text names the neighborhood, not just the object. "Finished hillside retaining wall in Palos Verdes Estates" does double duty — it's an accessibility requirement and a geographic signal, and most trades sites currently write neither.
Captions on the live page do the corroborating work review text usually does alone. "Completed March 2025 — Lunada Bay hillside, drainage and retaining wall" gives a human and a language model the same clear fact in one sentence.
Project pages get built by city, not by service type alone. A single "Our Work" page with forty mixed photos tells Google almost nothing about geographic range. Individual project entries — each with a location, a job type, and photos captioned to match — build a pattern a ranking system can actually read as "this company works across the South Bay," not just in the city on the letterhead. This is the same structural gap that shows up across home services sites generally: businesses that genuinely serve a 20-mile radius but only ever prove it in one zip code's worth of content.
GBP photo uploads get the same caption discipline as the website. Every photo pushed to the Business Profile should carry a specific, location-named caption — not because GBP reads EXIF, but because that caption is indexed text tied to a verified local entity, which is a far stronger signal than a stripped image file ever was.
None of this requires engineering. It requires someone to stop treating photo uploads as an afterthought and start treating them as one more place where the business's actual geographic footprint either gets documented or gets erased.
The Job Happened. Make the Internet Believe It.
A hillside job in Palos Verdes, a panel upgrade in Hermosa, a re-roof in Manhattan Beach — the work is real, timestamped, and sitting in a camera roll or a "Recent Projects" folder right now, proving nothing to anything that isn't a human scrolling by chance. The gap between doing the work and getting credit for the geographic range of that work isn't a Google penalty. It's a documentation habit nobody built.
Most South Bay trades businesses have more evidence of their actual service radius than their competitors do — they've just never structured it so a machine could read it. If your project photos are sitting in a generic gallery with no city names attached anywhere but your invoices, that's worth an actual conversation, not a plugin fix. We work with South Bay service businesses on exactly this kind of structural audit — the parts of the site that look fine to a homeowner and say nothing at all to the systems deciding who shows up first.