How to Quote Window Cleaning Jobs From Photos

Every drive-out estimate costs a window cleaning business real time — gas, a slot on the schedule, and often a wait for the homeowner to be home. A lot of that can be replaced with a good set of photos, if you know what to look for in them. Here’s how to actually do it.

What to ask the customer to photograph

The single biggest driver of estimate accuracy is photo coverage, not photo quality. A few blurry-but-complete shots beat one crisp close-up. Ask for:

• One photo of each elevation (front, back, and both sides) taken straight-on, not at an angle.
• Enough distance to fit the whole elevation in frame, including the roofline.
• A second, closer photo of any elevation with more than 6–8 windows, so individual panes are still distinguishable.
• Anything unusual called out separately — a sunroom, a bay window, a skylight, second-story windows with no visible ground access.

Counting windows and panes without miscounting

Count openings, not panes, first — an “opening” is one physical window unit, even if it’s divided into multiple panes by grilles or true divided lights. Then note the pane pattern separately (e.g. 6-over-6), since that affects labor time more than price-per-window alone. Two mistakes are worth watching for specifically:

• Reflections and glare can make a single large window look like two, or hide a window entirely against a dark tree line. Ask for a photo taken from a slightly different angle if a section looks ambiguous.
• Partially obscured windows — behind a parked car, a bush, or a porch roof — are the most common source of a low estimate that turns into a surprise on-site. Flag these rather than guessing.

What else changes the price

Window count gets you most of the way to an estimate, but a few other things consistently matter more than people expect when pricing from photos alone:

• Stories and access. A second-story window over a flat, walkable roof is a different job than one over a steep pitch or a pool.
• Screens and tracks. Ask whether screens should come out and get cleaned separately, and whether tracks look like they’ve been neglected — both are easy to miss from a photo taken a few feet back.
• Hard water staining. Visible etching or mineral buildup usually means a different (often billed separately) treatment, not just a standard clean.
• General condition. Years of buildup takes longer to clean than routine maintenance, even at the same window count.

When to still schedule a site visit

Photo-based estimating is a filter, not a replacement for judgment. Keep visiting in person when photos don’t clearly show enough of the property to be confident, when a customer describes something the photos don’t cover (interior cleaning, an unusual window type, commercial-scale glass), or when the estimate is large enough that getting it wrong is expensive either way. A tool that tells you when its own confidence is low — rather than guessing silently — makes this call a lot easier.

Where AI-assisted tools fit in

This is exactly the workflow Tallyvis’s AI estimator is built around: a customer uploads photos on your own website, the AI proposes a window count and job characteristics, you (or the customer) confirm anything it flagged as uncertain, and the price comes from your own pricing rules — never invented by the AI. You can try the live demo estimator to see what that confirmation step actually looks like.

FAQ

How many photos does a customer usually need to send?

Four to six is typical for an average single-family home — one per elevation, plus a closer shot of any busy elevation. More photos rarely hurt; fewer photos is the more common problem.

Should I always show a price range instead of an exact number?

When photo coverage is genuinely incomplete, a range (or a note that the number needs confirmation) is more honest than a precise-looking figure you’re not confident in — and it sets the right expectation before you ever show up.