Skip to main content

HOME SERVICES MARKETING · September 2026 · ~10 min read

Photos of your work: how they affect the quote request

Photos of finished jobs answer the question a homeowner cannot ask directly: will my house look worse after you leave. Stock images answer nothing, because the visitor knows they are stock. The photos that change quote requests are your own work, in houses that look like theirs, with the mess cleaned up.

Contractors treat photos as decoration. An agency asks for images, somebody sends twelve phone pictures of a truck and a crew, and the rest gets filled with stock.

That is a missed opportunity, because in this business the photo is the proof. A homeowner cannot evaluate your soldering. They can absolutely evaluate whether the finished install looks tidy.

01

What is the homeowner actually looking for?

Three things, and none of them are what contractors photograph.

Does this look like my house. A photo of a commercial rooftop unit tells a homeowner in a 1950s ranch nothing. Same trade, different world. Match the housing stock in your market.

Did they clean up. This is the fear. Not the quality of the work, the state of the hallway afterward. A finished job photo with drop cloths gone, tools gone and the floor swept does more than any claim you could write.

Is it real. Homeowners recognize stock photography instantly, and the recognition is expensive. A stock photo does not read as neutral. It reads as a company with no finished work to show.

The best photo on most contractor sites is the one a tech took with a phone before packing up, straight on, with the work visible and the site clean. Not the professionally lit one.

02

Before and after, or just after?

Both, and the before is doing more work than people expect.

A before photo establishes the problem. A homeowner staring at a corroded shutoff valve does not know what they are looking at until they see the same valve in your before shot, followed by the after. Now they can identify their own situation and they know you have handled it.

That identification is what converts. The quote request comes from "that is my exact problem," not from "these people seem good."

Pair the photos with a caption naming the job and the town. Three sentences. What was wrong, what you did, roughly how long it took.

If you run multiple trades, keep the photo sets separate. A plumbing gallery mixed with roofing photos reads as a handyman operation, which is why multi-trade companies need a deliberate website structure rather than one combined portfolio.

03

Where do the photos belong?

Next to the ask, not on a gallery page.

The gallery page is where photos go to be ignored. Almost nobody clicks "Our Work" from the menu. The visitor is on your water heater page deciding whether to submit a form, and that is where the water heater photos need to be, directly above it.

Sequence matters. Problem photo, solution photo, caption, form. Then repeat for the next job type.

Two other elements belong in the same stretch of page. Your license number and insurance status, because licensing and insurance are the trust signals that change who submits, and payment options, because a homeowner looking at a five figure replacement is doing arithmetic. Where financing goes on the site determines whether they ask for a price or close the tab.

Proof, credentials, payment, ask. In that order, on the same screen.

04

Do photos slow the page down enough to matter?

Yes, and this is where a good idea quietly becomes an expensive one.

Google's own research on mobile page speed, using a model it says predicted bounce and conversion with 90% accuracy, states that as page load time goes from one second to 10 seconds, the probability of a mobile site visitor bouncing increases 123%. The same article reports that as the number of elements on a page goes from 400 to 6,000, the probability of conversion drops 95%.

The image specific numbers are in the supporting analysis of 11 million mobile domains: 79% of pages were over 1MB, 53% over 2MB and 23% over 4MB, and 25% of pages could save more than 250KB by compressing images and text alone. Uncompressed job photos are the single most common reason a contractor page is in the 4MB group.

Google's current Core Web Vitals threshold for Largest Contentful Paint is 2.5 seconds or better, assessed at the 75th percentile of visits. On a photo heavy service page, the largest contentful element usually is a photo. The image you added to win the job is frequently the one deciding whether the page renders in time to be seen.

None of that means fewer photos. It means compressed photos, sized for the slot they sit in, loaded below the fold where they belong.

05

What does that look like in arithmetic?

Worked example

Here is a full pass on one page. Rerun it with your own analytics.

Step one, the baseline. Your water heater page gets 620 visits a month and produces 18 quote requests. That is 2.9%derived.

Step two, the weight. Fourteen job photos straight off a phone at roughly 0.75MB each is 10.5MB, plus 0.9MB for everything else. The page is 11.4MB.

Step three, the fix. Resize and compress each photo to about 0.15MB. Fourteen photos is now 2.1MB, so the page is 3.0MB. That is a 74%derived reduction and nobody will see a quality difference on a phone.

Step four, the rebuild. While you are in there, replace six stock images with before and after pairs from real jobs, each with a three sentence caption, and move the form directly under the third pair.

Step five, remeasure at ninety days. 640 visits, 26 quote requests, 4.1%derived.

Step six, be honest about what you just learned. You made two changes and measured once, so you cannot say which one produced the eight extra requests. At 18 conversions a month, splitting traffic into two variants leaves nine per variant, which will not separate anything in a reasonable window. Report it as a before and after with both changes named, not as a controlled test. That is the honest version, and it is still worth doing.

Step seven, value it at the end of the funnel. Eight extra quote requests are not eight jobs. Apply your own close rate and your own average ticket, from your own field service software, and report cost per booked job rather than cost per lead. If you close four of the eight, the page paid for an afternoon of work many times over. If you close none, the photos are not your problem and your quoting is.

06

Is there a page speed number you should refuse?

Yes, and it is probably the one in the last proposal you read.

The claim in circulation is that "as page load time goes from one to three seconds, the probability of bouncing increases 32%," usually attributed to Google. It does not appear in either primary Google document. The 2017 edition of Google's benchmark article says one second to seven seconds and 113%. The 2018 update, which is the live version, says one second to ten seconds and 123%. There is no one to three second, 32% finding in either.

There is a second trap in the same family. Because both Google editions exist, you will see 113% and 123% quoted side by side as though they were two independent studies confirming each other. They are one finding, rerun with a different scale.

So: cite the 123% figure, name the year, and treat any page speed table without a document behind it as decoration. Speed is worth money. The specific numbers people sell it with are frequently invented.

For what a controlled measurement looks like, Google's own study with Deloitte and Fifty-Five, covering 37 brands and about 30 million sessions in October and November 2019, found a 0.1 second improvement in mobile load time associated with retail conversions up 8.4% and average order value up 9.2%, and for the six lead generation brands in the sample, bounce rate on information pages improved 8.3%. That is Google measuring its own thesis, on brands much larger than yours, seven years ago. Directionally useful. Not a promise about your water heater page.

07

How do you collect photos without the system dying?

Make it one person's job at one moment, or it will not happen.

The universal failure is asking every tech to photograph every job. Two weeks in, nobody does. Techs are paid to finish work, not to build a content library, and any system depending on ten people remembering something dies.

What works is narrower. Pick one job type per month. Tell the lead tech on that crew you need three photos from any two jobs. Before, after, and one wide shot of the clean site. That is six photos a month, seventy a year.

Ask permission for exteriors and never publish an address. Interior shots of a mechanical room are low risk. A front elevation with a house number is not.

Then use them everywhere: profile posts, the service page, the city page, ad creative. One caution on the profile. Whitespark's 2026 report added presence of AI generated photos or videos on the profile to its list of negative and suspension risk factors. Generated or heavily synthetic images are not a shortcut around an empty camera roll. Use real photos or use none.

A photo library is one of the few marketing assets that gets more valuable the longer you keep it, unlike ad spend, which resets monthly and needs watching so the budget does not run out mid month.

08

What to do this week

Open your website and count the photos that are actually yours. Not the truck, not the crew, finished work. If the number is under five, that is the project.

Go through your techs' camera rolls. There are already dozens of job photos on those phones, taken for warranty or documentation. Ask for them. Fastest source you have and it costs nothing.

Run your highest value service page through any page speed tool and write down the current size and Largest Contentful Paint. Then do steps two through five above.

And if better proof produces more calls, know what that costs downstream. How call volume and overage pricing work on AI phone systems is worth understanding before your volume changes, not after the invoice.

Be honest with yourself

When you do not need this

If your work is invisible, photos will not help much. Slab leak repair, duct sealing and water treatment do not photograph in a way a homeowner can read. Use diagrams or short video instead, or lean on reviews.

If you do emergency work almost exclusively, nobody browses your portfolio at 11pm. Your first screen and your phone answering matter far more.

And if you are booked out and turning away work, adding proof to increase demand is the wrong move. Raise prices instead.

Sources

Related reading

12

Questions about your job photos?

Email me at eric@seod.com with two or three photos from a recent job and the service page they would go on. I will tell you which one to lead with, what caption to put under it, which to leave out, and what the page currently weighs. Quick answer, nothing attached, and I will be blunt if none of them are usable.

More on converting contractor traffic sits in the home services library.

Call Eric Email Eric