Nobody wakes up wanting a data analyst. They want Monday's report to stop taking all of Monday. The first person who makes that specific problem disappear for a fixed price becomes the analyst the company keeps.

That is the answer to how to find clients for data analytics: skip companies that already have data teams, find mid-size businesses that run on spreadsheets, spot one report they obviously build by hand, and offer to automate just that report. This article walks through how to hunt for the report, what to charge, the first email and how a one-off project becomes monthly work.

Who pays for analytics (hint: not data teams)

Companies with a data team already have analysts, and they are hiring full-time people, not freelancers. Your buyers are the opposite: businesses with 20 to 200 employees that have outgrown a spreadsheet and have never hired for it. A distributor with three warehouses. A multi-location clinic or restaurant group. A construction firm tracking job costs. A property manager with 600 units. A regional retailer with several stores.

The buyer is usually one of three people. The owner, who wants to see the business on one page. The operations manager, who builds the reports and hates it. Or the CFO or controller, who owns month-end and would pay to cut it from five days to two. A company with a trigger is easier: a new location, a new system, a missed forecast, an acquisition or a bank asking for cleaner numbers.

Companies with messy spreadsheets and money riding on them

The best prospects are not the ones with the most data. They are the ones where a wrong number costs real money. Inventory counts that decide purchasing. Job costs that decide whether a bid made money. Staffing hours that decide payroll. Patient or customer volume that decides schedules. If the report feeds a decision made every week, someone is already spending hours on it, and that is your budget.

Owen Pak, a freelance analyst in Minneapolis, targets mid-size distributors and contractors within 60 miles. He is not selling "analytics." He is selling the end of one weekly chore.

The report-hunting method

You can find these reports from the outside. Four signals work.

  1. Job postings. Search for roles that mention "prepare weekly reports in Excel," "compile data from multiple systems," "reconcile," "update spreadsheets" or "month-end reporting." A company hiring an assistant to build a report by hand has told you exactly what to automate. The hiring manager's name is your first email.
  2. Job titles. A company with an "operations coordinator" or "reporting analyst" but no data tools in the posting runs on spreadsheets.
  3. Public numbers that look hand-made. A distributor's website with an "in stock" page that updates on Mondays, or a property manager's monthly owner statements as PDFs, usually come from somebody's spreadsheet.
  4. A direct question. In every conversation, ask what report people dread. A tool or a person who answers that question once is worth more than any research.

For each prospect write the company, the person, and the one report you guess they build by hand, with the evidence in a few words ("posted for 'reporting coordinator,' lists Excel and pulling data from three systems"). Do 40. Rocketship can search by title and company size (operations manager, controller, owner), and preview 25 names before spending a credit; the report guess is yours. For the list-building side, how a freelance consultant gets clients covers the general case.

Offering one dashboard at a fixed price

Owen's offer is a single page: one dashboard, up to three data sources (say Excel, QuickBooks and a point-of-sale export), a refresh schedule, one round of revisions, a 30-minute walkthrough for the team and 30 days of fixes. The price is $2,400 and the timeline is about two weeks. Fixed scope is the point. A vague "analytics project" scares owners; a named report with a price does not.

Do the owner's arithmetic with them. If the ops manager spends 12 hours building the month-end report and her time costs the company about $40 an hour, that is $480 a month, or $5,760 a year. A $2,400 dashboard pays back in five months, and that ignores errors and the days the numbers arrive late. Say it plainly, and ask them to correct your assumptions.

A first email that names a report they probably build by hand

Subject: The Monday inventory report at Birchwood Supply

Hi Lorraine, I saw Birchwood Supply is hiring a reporting coordinator to "compile weekly inventory and order data in Excel." My guess is someone builds that report by hand every Monday from the order system and the warehouse counts. I build dashboards for businesses like yours: one report, fixed price, connected to your files so it updates on its own. If I'm wrong about the report, tell me which one it is. I'd like to mock up a sample using made-up numbers so you can see it.

Owen Pak, Pak Analytics, 615 Hennepin Avenue, Minneapolis. Reply "no thanks" and I won't write again.

The hedge ("my guess is") is honest, and it invites a correction, which starts a conversation. Include your physical address and an opt-out line, as US email rules require. For making each first line specific at low volume, see personalisation at low volume.

Following up with a sample dashboard

If she says "send it," or even if she does not reply in four days, send a one-page picture of the dashboard filled with invented data laid out like her report: inventory by warehouse, orders waiting, items below reorder level. Never use real numbers you do not have. Label it "sample, made-up numbers." Write one sentence: "This replaces the Monday report. It updates when your files update."

On day 10, send a short note with a different angle: "If it's easier, I can do a 20-minute call to hear how the report is built today." Then stop. More on spacing in cold email follow-up sequence. Your prospects' sheets are not wrong, just overworked. CRM vs spreadsheet shows the point where a sheet stops being enough on the sales side, and the same threshold logic helps you explain it for reports.

Moving from project to monthly support

At delivery, ask: "Which other report does someone here build by hand?" Almost every company has three. Quote the next one at a lower price, because the data connection exists: Owen charges $1,200 to $1,800 for a second dashboard on the same sources. And offer a monthly plan: $400 a month for four hours, covering fixes when a source changes, small new metrics and a quarterly review of what the numbers say.

Monthly support is where freelance analytics becomes a business. Ten clients on $400 is $4,000 a month before any new project. To sell it, point at the thing that breaks: a renamed column, a new product line, a new location. The dashboard needs someone who understands it. Related reading on selling technical work to businesses without technical staff: how to get clients for an AI automation agency and how to get clients as a fractional CFO, since the CFO is often your champion. If you also handle the systems the data comes from, how to get clients for an IT company covers the managed-services side. For a nearby craft that uses the same free-teardown idea, see how to find clients for UI UX design.

Finding 40 companies, guessing 40 reports and following up on each is real work. Rocketship finds the buyers, emails them from your own Gmail, follows up, and books the meeting on its own. Plans that get you customers start at $24.99 a month (Launch), and every monthly plan starts with a 7-day free trial, card required. The AI workers handle the sending and follow-up so your time goes to the reports.