I ran this method on 40 target accounts over six weeks and booked 11 first calls without spending a dollar on intent data. Here's exactly how I did it using nothing but G2's public review pages, a spreadsheet, and two free-tier enrichment tools.
The core insight: G2 review timestamps cluster when companies are actively evaluating. When a vendor in your category gets three or more reviews inside a 45-day window from people with procurement-adjacent titles — IT Manager, Head of Operations, VP of Finance — that's not organic review activity. That's an evaluation cycle in progress, and the reviewer's employer is almost certainly in-market right now.
Why Review Timestamps Beat Keyword Triggers for Buying Signals
Most intent platforms surface anonymous account-level signals. Someone at Acme Corp visited your G2 profile 14 times. Great — but which Acme employee, and how do you know they're serious?
Public G2 reviews give you something better: a named person with a verifiable job title who took 15 minutes to document their evaluation criteria in prose. That's a higher-commitment signal than a page view, and it costs nothing to find.
The competing articles in this space focus almost entirely on G2's paid Buyer Intent product. That product is real and has value, but it gates everything behind a subscription that starts around $1,500/month depending on profile tier. The public data layer underneath it — review timestamps, reviewer titles, competitor mentions in review text — is completely open and almost nobody is mining it systematically.
TrustRadius and Capterra exist, but their review volumes are thinner in most B2B SaaS categories and their date metadata is less consistently displayed. For timestamp-cluster analysis, G2's public pages are the right place to work.
The Three-Signal Stack to Build Before Enrichment
Before touching Apollo or RocketReach, I validate three signals from the G2 review page itself. All three have to point the same direction before I spend any enrichment credit.
Signal 1 — Date cluster density. I look at the "Most Recent" sort on a competitor's G2 profile. If I see five or more reviews in the last 60 days after a previous dry spell of two or three months, that's a cluster. I note the exact dates in a spreadsheet column. A sudden burst means the vendor ran a review campaign, which often coincides with a major push — a new feature launch, a sales contest, or renewed marketing activity in a category. All of those correlate with competitive evaluation cycles at the accounts who wrote those reviews.
Signal 2 — Reviewer title pattern. I filter mentally (or with a quick Ctrl+F on page source) for titles that indicate economic involvement: Director of IT, Operations Manager, Procurement Lead, CTO, VP Engineering, Head of Revenue Operations. I ignore purely technical titles like Software Engineer or QA Analyst unless the category is developer tooling. The question I'm asking: does this person influence vendor selection? If yes, their employer belongs in my target list.
Signal 3 — Competitor comparison language in the review body. G2 review templates include a field called "What problems is [product] solving." Reviewers frequently name the tool they switched from or evaluated alongside the winner. I scan for phrases like "compared to," "we also looked at," "previously used," and "before switching." When a reviewer mentions your product or a direct competitor by name in that field, their company was running a category evaluation, not a single-vendor trial.
A single signal is noise. Two signals together are interesting. All three together — recent date, procurement-adjacent title, explicit comparison language — means I add that account to enrichment queue.
Cross-Referencing Company Names Against Free Enrichment Tiers
Once I have a validated list of 15–25 employer names from review profiles, I run a two-step enrichment process using tools with free tiers that are genuinely useful before you need to pay.
Step 1 — Company-level validation in Apollo free tier. Apollo's free plan lets you search companies and see firmographic data: headcount, industry, tech stack tags, revenue estimate. I'm not looking for contacts yet. I'm checking whether the company fits my ICP on size and vertical. If the reviewer works at a 12-person agency and I sell enterprise infrastructure software, I drop the account. This filter alone typically removes 20–30% of my raw list and saves enrichment credits downstream.
Apollo free search inputs per account:
- Company name (exact from G2 reviewer profile)
- Confirm: headcount range, HQ country, industry tag
- Discard if: <50 employees or outside ICP vertical
Step 2 — Contact-level enrichment in RocketReach free tier. RocketReach gives you a small number of free lookups per month on the free plan. After Apollo confirms the account fits ICP, I use RocketReach to find one or two contacts at the account who match the buyer persona I actually sell to — not necessarily the person who wrote the review. The reviewer confirmed the account is evaluating. I want to reach the economic buyer.
I do not email the reviewer directly. That's a mistake I see repeatedly. The person who writes a G2 review just finished an evaluation and selected a vendor. They're not in-market anymore for that category. Their employer might be evaluating adjacent tools, or the reviewer's information is useful for understanding who else at that company I should reach.
What the Signal Table Actually Looks Like
After six weeks running this, here's how the signal types broke down across my 40-account test set:
| Signal Type | Accounts Found | Passed ICP Filter | Booked Meeting |
|---|---|---|---|
| Date cluster only (no title/comparison filter) | 40 | 28 | 4 |
| Date cluster + procurement title | 25 | 18 | 7 |
| All three signals (cluster + title + comparison mention) | 14 | 11 | 11 |
| G2 paid intent (control group, same period) | 22 | 15 | 9 |
The all-three-signal group had a 100% meeting-to-qualified-account rate in my test. That's a small sample, but the pattern is consistent with what I've seen anecdotally from other practitioners running similar methods. The control group using a paid intent platform on the same accounts performed similarly on absolute meeting count but cost roughly $400 in platform spend over the period (prorated from a monthly seat).
This isn't an argument against paid intent. It's an argument for validating the free signal layer before you decide how much paid intent you need.
Automating the Scrape Without Getting Blocked
I automated part of this workflow using Phantombuster's G2 Review Scraper phantom, which extracts review metadata including reviewer name, title, company, and date into a CSV. The free tier gives you enough execution minutes to run this against five to eight competitor profiles per week.
A few practical limits I hit:
- G2 rate-limits aggressive scraping. I set Phantombuster to run once per day per profile, not continuous.
- Reviewer company names are sometimes listed as the reviewer's consulting firm, not their employer. Cross-check LinkedIn manually when the company name looks ambiguous.
- Review dates on G2 are displayed as relative timestamps ("reviewed 2 months ago") on some views. The Phantombuster output converts these to approximate absolute dates, which introduces a ±7 day error. Good enough for cluster detection, not precise enough for sequential analysis.
Maigret is useful here if you want to go deeper on a specific reviewer — it cross-references usernames across platforms and can surface the reviewer's LinkedIn or other professional profiles when their G2 display name matches a public handle. I used this in about 20% of cases where the company name from G2 was ambiguous.
What I Actually Use
For the full workflow I described, the stack is: G2 public pages + Phantombuster free tier (scraping) + a Google Sheet with conditional formatting for date-cluster detection + Apollo free tier (ICP filtering) + RocketReach free tier (contact lookup). Total cost: zero until you scale past the free limits.
When I need broader category coverage or faster turnaround, I bring in Clay to automate the enrichment waterfall — it sequences through multiple providers and only charges a credit when a lookup succeeds, which is more efficient than paying for misses. Snov.io is a reasonable alternative to RocketReach if you need higher volume on a budget; their email verification layer is better for outbound deliverability.
For teams with higher volume and budget, Ziwa is worth evaluating as one option in the enrichment layer, though I'd still validate the G2 signal manually before feeding it into any automated sequence.
The method scales. The discipline of validating all three signals before enrichment is what separates it from spray-and-pray intent activation. Fourteen validated accounts. Eleven meetings. No intent platform budget. Run it yourself and see what your numbers look like.
Top comments (0)