Top window cleaning software, rated on evidence rather than hype

What to take away

This is general business guidance. Data protection, employment and tax obligations vary by jurisdiction, so take qualified advice on how they apply to whatever system you adopt.

The shortlist, ranked

Ranking here means one thing: how much of the job record the software captures without a workaround. A tool that schedules well but cannot tell you time on site loses to one that does both.

The order below uses the public plan pages from Jobber, Housecall Pro, ServiceTitan, GorillaDesk, ZenMaid, Launch27 and BookingKoops, plus vendor documentation, not demo accounts. The named sources behind the rest of the article each carry a direct link as well.

Those are the Internal Revenue Service for recordkeeping, the National Institute of Standards and Technology and the Cybersecurity and Infrastructure Security Agency for data security. The Occupational Safety and Health Administration covers fall protection.

Jobber is the common first paid system for a window cleaning company because quoting, scheduling and invoicing sit in one account, and it exports.

Housecall Pro leans residential, with online booking and card payments that shorten days to collect.

ServiceTitan is heavier and priced by quote, which suits operators running several crews who need dispatch and reporting.

GorillaDesk and ZenMaid are the low-cost end. Both handle recurring visits and invoicing for a small crew without a per-user bill that grows every time you hire.

Launch27 and BookingKoops are booking engines first. They take the web enquiry, offer a slot and turn it into a scheduled job, which is what a route business needs from its front end.

Confirm current pricing on each vendor's own page before you budget. Rates change, and a figure quoted here would be stale by the time you read it.

What the system has to do

Strip the marketing away and a window cleaning company needs the six measures listed below recorded, plus one export requirement.

If a system does the first six and fails the seventh, treat that as a serious problem rather than a detail. Your history is what prices next year's work.

The six measures that matter

1. Revenue per working day. Total revenue divided by days actually worked. The headline number, and the one that shows seasonality honestly.

2. On-site hours as a share of paid hours. The route efficiency measure. It moves when density changes, when the schedule is built badly, or when a crew is spending the day driving.

3. Contribution per productive hour. Revenue minus materials, divided by all crew hours including travel and setup. The measure that ranks your work.

4. Recurring share of revenue. What proportion comes from signed frequencies rather than one-off jobs. The best single indicator of whether you own a business or a job list.

5. Rework rate. Return visits as a share of jobs completed. A full cost with no revenue attached, so it is worth watching closely.

6. Days to collect. From invoice to money in the account. Profit and cash are different, and this is the measure that explains why a profitable month can still be tight.

Six is enough. Companies that track twenty measures act on none of them.

Setting thresholds you will actually use

A measure only becomes management when it has a threshold attached.

Nobody can give you those thresholds, because they depend on your service mix, your geography and your season. What you can do is this: run all six for three months, take your own median as the baseline, and set a threshold at a movement large enough to be worth a decision.

Worked example with round numbers. A two-van residential company bills $280,000 over 200 working days, so revenue per working day is $1,400. Crews are paid for 3,600 hours and are on site for 2,400 of them, so on-site share is about 67 percent. If materials cost $28,000, contribution per productive hour is about $70.

Rework of 4 percent means 40 return visits across 1,000 jobs. Days to collect of 30 means the cash from a job lands about a month after the invoice. These figures illustrate the arithmetic rather than set targets, and your own three-month median replaces them.

Then write, for each measure, the action that follows. Rework above your line means a training or scheduling review. On-site share below your line means the route is wrong. Days to collect above your line means the invoicing routine needs attention, not the customers.

The records the measures come from

Every one of these numbers is built from per-job capture: time on site, travel, materials, revenue, and whether a return visit happened. Capture it at the job. Reconstructed timings are guesses wearing a spreadsheet.

The Internal Revenue Service recordkeeping guidance sets out the standard a system has to reach before it can show income and expense clearly, which is the same standard that makes the six measures trustworthy a year later.

Your system holds other people's data

Once you have customers, you are holding sensitive details. These include names, addresses, phone numbers and access arrangements. They also include gate codes and photos inside homes. That is a genuine responsibility, and window cleaning companies rarely think of themselves as data holders.

The small business cybersecurity quick start guides from the National Institute of Standards and Technology set out the basics. They are in a form a small company can actually apply. The small and medium business material from the Cybersecurity and Infrastructure Security Agency covers the same ground. It offers practical steps.

The short version for this trade: separate accounts per person, multi-factor authentication on anything holding customer records, backups you have actually tested restoring, and removing access the day somebody leaves. Access codes in particular should live in the system with restricted visibility, not in a group chat.

Choosing without being sold to: a checklist

The app design and the feature list matter far less than these checks.

From measures to decisions

Contribution per hour and revenue per working day feed straight into the pricing and profit decisions, because a price built without them is a guess. On-site share and rework are operational, and the fixes live in the operations and workflow guide rather than in a spreadsheet.

Rework that concentrates on one crew is a training signal, which is the subject of hiring and training window cleaners. Recurring share is a selling signal and belongs with the marketing and growth guide, since it measures whether first visits are being converted into agreements.

Two constraints sit around the system. Some of what you store, particularly training records, equipment inspections and insurance certificates, exists because of the licensing and compliance requirements that apply where you work. And the practical usefulness of a job record depends on the crew being able to complete it on site with the equipment and devices they carry.

Moving in without losing a season

Migration is where good software decisions go wrong. Do it in the quiet part of your year, and in this order.

Run the old and new systems in parallel for one full cycle. It is duplicated effort for a few weeks and it is much cheaper than discovering in month two that a frequency was never created.

Safety records belong in the system too

Equipment inspections, training sign-offs and access method decisions are records, and records that live in a drawer are records nobody can produce.

The fall protection guidance from the Occupational Safety and Health Administration frames protection as planned and verifiable, and the practical version of verifiable is a dated entry somebody can find. Keeping inspections and sign-offs in the same system as the schedule means the crew lead can check them without asking anyone.

Common questions

Do I need software at all as a one-person company?

Not on day one. A notebook works at ten accounts and becomes dangerous at thirty. The right moment to move is while migration is still small, not when the notebook fails.

What is the most common mistake in choosing a system?

Choosing on features and discovering later that the job record does not capture time on site. Without that field, most of the measures above are unavailable and the pricing review has nothing to work from.

Should the crew enter data on their phones?

Yes, at the job. It takes less time than reconstructing it, and photographs attached to the job at the moment they are taken are worth far more than photographs found later in a camera roll.

How often should I look at the six measures?

Monthly for all six, and weekly for on-site share and rework while a route or a crew is new. Looking daily produces noise rather than management.