Module 5: Research and Data
Research and data work is where a marketing VA quietly becomes indispensable. Anyone can find “some leads.” You deliver lists built exactly to spec, a CRM that tells the truth, and reports a client can read in ninety seconds. Four lessons, one standard: accurate beats fast, every single time.
Prospect research: build to the spec
Every research task starts with a spec, and if the client hands you a vague one, your first deliverable is two questions: who counts, and who is an automatic no? A workable spec names the target (industry, business size, location, the person's role), the exact fields per row (name, role, company, website, business email, profile URL, one qualifying note), the agreed sources (company websites, LinkedIn, industry directories), the quantity, and the deadline. What the list will be used for — outreach, event invitations, partnership scouting — decides which fields matter, so ask that too.
Then work inside the sheet, never in your head. One row per prospect, named columns, and a source note for every row: the page where you verified that person actually holds that role. Dedupe as you go against the client's existing list and CRM — delivering 40 people they already have is worse than delivering 25 new ones. The discipline that defines this work: no padding. Eighty verified rows beat a hundred rows with twenty guesses stirred in. A person who doesn't meet the spec gets excluded, and a field you could not verify gets marked unverified — it never gets invented.
Protect yourself and the timeline with a sample batch. Before researching 300 rows, deliver the first 10 and get a yes on format and fit. A misunderstood spec caught at row 10 costs an email; caught at row 300 it costs the engagement. Clients also read that early sample as a signal: this VA checks before charging ahead. That signal is worth more than the rows.
The spec says: owners of dental practices in Texas, personal business email required. Row 47 turns out to be an office manager whose only email is a Gmail address. What belongs in the sheet?
List building and verification basics
A list is a promise the client makes to their email platform: these are real people who belong here. Build it from sources in order of trust — the client's own records first (past customers, event sign-ups, inquiry forms), then public business sources like company websites and directories. Every row carries a source note, because six months from now someone will ask where a contact came from, and “the list” is not an answer.
Verification is a series of small checks, done by hand, row by row. Does the email's domain match the company on the row? Does the company website or profile still show this person in this role, or did they leave last year? Is the address a person or a role inbox like info@ or sales@ — fine if the spec allows it, wrong if it asked for people. Is the formatting clean: consistent capitalization in names, one phone format throughout, state names handled the same way on every row? Clean as you build, not in a panic the night before the import.
Understand what is at stake. When a client emails a dirty list, messages bounce and land in spam, and the damage attaches to their domain — the same domain their invoices and customer replies come from. Your verification work is what stands between them and that outcome, so price it and describe it that way in your updates: not “cleaned a spreadsheet,” but “verified 240 contacts against company sources; removed 31 stale or mismatched rows before import.” That sentence is the job, stated in the language clients pay for.
A new client hands you a purchased list of 5,000 emails and asks you to import it into their email platform and schedule a campaign for tomorrow. What is the professional move?
CRM hygiene: one true record
A CRM exists to answer one question instantly: what is true about this relationship right now? When it holds three records for the same person, two tag styles, and a pipeline stage from four months ago, it answers nothing — and every follow-up, report, and campaign built on it inherits the error. Hygiene is the unglamorous work of keeping that record true, and clients feel its absence every day even when they can't name it.
The three habits that do most of the work. Merge duplicates, don't just delete them: keep the record with the richest history, fold the other's notes and activity into it, then remove the empty shell — deleting outright can erase the one conversation that mattered. Tags come from an approved list only: propose a small taxonomy (source, status, interest) and get the client to sign off, because free-typed tags decay into “Webinar,” “webinar,” and “WEBINAR-2025” within a month. Stages match reality: a contact who replied sits in a different stage than one who booked a call, and you update the stage as part of the task that changed it — not in a monthly cleanup marathon.
Notice where this leads. The VA who visibly keeps a CRM true is the first person offered the setup and cleanup projects behind it — and tech and CRM work sits in the $40–80/hr specialty band, against $20–40 for general admin. Hygiene done reliably, and reported plainly (“merged 14 duplicate records, retired 6 stale stages this week”), is the on-ramp to that band. You are not tidying a database. You are auditioning for the premium tier of the trade, in public, every week.
You find the same prospect three times in the CRM: one record with the full email history, one with a phone number, one sitting in the wrong pipeline stage — and the tags are spelled two different ways. What's the right sequence?
Reporting support: same numbers, same way, every time
Most reporting support is pipeline work, not analysis: pull the agreed metrics, from the agreed sources, on the agreed day, into the agreed template. That sounds dull until you understand what it buys — comparability. A client can only see a trend if this month's “email revenue” was measured exactly the way last month's was. The moment a VA changes a source, a date range, or a definition without saying so, the chart starts lying and nobody knows when it started.
So define everything once, in writing, in the reporting SOP: each metric, where it comes from, which account, which date range, pulled on which day. If a platform later restates a number — it happens — note the change in the report instead of silently rewriting history. Send the report in the client's reading order: headline numbers first, change versus last period, then one line of context each. You are not expected to write strategy, but you are expected to flag missing data (“social stats delayed on the platform's side; updating Thursday”) rather than leaving a gap for the client to discover.
One habit separates trusted VAs from copy-paste VAs: sanity-check before you send. If clicks fell off a cliff or sales tripled overnight, suspect the pipeline before you suspect the business — a broken tracking tag, a date range that caught a holiday weekend, a dashboard filter left on. Report the anomaly with your best explanation (“shows −90%; the range included the platform outage day, recheck scheduled”), never raw and unexplained. And the mirror rule: never pick a flattering date range or round a number into a better story. Honest ugly beats polished wrong, because the client is making spending decisions on your sheet.
This week's report shows email clicks down about 90% from last week. The client reads the report in an hour and makes budget calls from it. What do you do first?
What you can now do
- Sharpen a vague research ask into a working spec — who counts, who is an automatic no, the exact fields per row — before you build a single row.
- Build a lead list from sources in order of trust, client's own records first, with a source note on every row and verification before it ships.
- Keep a CRM at one true record per person, and pull reports the same numbers, same way, every time — because a trend only exists if the measuring never changes.