The short answer
Get read access to the seller's Google Analytics property and Search Console site; screenshots are not proof. Then compare total sessions with Search Console clicks month by month, find out which sources drove any jump, and confirm the property's top pages match the site for sale. A sudden change in that ratio needs an explanation before you sign.
- Direct, read-only access beats any export the seller prepares
- Look for a stable ratio over time, not for numbers that match exactly
- Growth from one referrer or an ad budget can disappear after the sale
Red flags that end a deal early
Most bad listings give themselves away within an hour. One third-party buyer's audit guide from Organic Arbitrage says two or more red flags demand explanations, and clear fraud is a reason to walk.
Stop and ask the seller when you see
- Proof only as pictures. Screenshots, PDFs or a recorded walkthrough, and excuses instead of access.
- A ratio that lurches. Total sessions divided by Search Console clicks holds steady for months, then jumps by a third or more in one month.
- A new referrer at the top. One outside domain that sent nothing a few months ago now sends a large slice of sessions.
- Ad spend nobody mentioned. The listing says all traffic is organic, but the site's Google Ads account spent money in the last 90 days.
- Pages that don't line up. The pages with the most search clicks barely appear among the property's top landing pages.
- A history with holes. The property's data starts recently, or shows days with zero sessions mid-season.
- Heavy direct traffic. The Organic Arbitrage guide treats Direct above about 30% of sessions as a warning sign worth a question.
- A hockey stick on a young site. As the Investors Club due diligence checklist (third party) puts it, "A hockey-stick traffic chart is the easiest way to get fooled."
None of these proves fraud alone. A press mention can create a new referrer. But each has a checkable explanation, and a seller who knows their site gives it in a sentence.
What to request from a seller, and what to refuse
Ask for access to the source, under your own Google login, with the lowest role that still shows the data. The Investors Club checklist says to "Demand live, read-only access to dashboards or raw, unedited CSV exports straight from the source," and warns that a recorded video can be staged.
| What the seller offers | Proves the numbers? | What to do with it |
|---|---|---|
| Report screenshots | No, an image is easy to edit | Thank them, then ask for GA4 access |
| PDF or CSV export | No, the seller chose the rows and dates | Pull the same range yourself and compare |
| Live screen share | Partly, the seller still drives | Better fit for one job: showing the GA4 data stream |
| GA4 Viewer access | Yes, you read the property directly | Ask at property level, under your own login |
| Search Console restricted user | Yes, clicks come from Google | Ask for the Domain property if one exists |
| Google Ads read-only | Yes, for spend history | Ask even when the seller says ads never ran |
Put three role names in your written request, using Google's own labels as of November 2026: Viewer in GA4, restricted user in Search Console and Read-only in Google Ads; the access guide covers what each can see and the clicks the seller needs.
Politely refuse two things: a "copy" property set up for buyers, and a URL-prefix Search Console property when the seller also owns a Domain property. The property check below explains why.
Cross-checking Analytics against Search Console
This is the most useful test a buyer can run, because the two numbers come from different systems. Analytics counts sessions with a tag the seller controls, so anyone with edit rights can add a second tag or point junk visits at it. Search Console counts clicks from Google's results pages. Inflating the first is easy. Inflating the second is much harder, because Google records those clicks itself.
They will never match, and Google says so. Its Search Console troubleshooting page notes that tools like Google Analytics "track traffic only from users who have enabled JavaScript," and that Search Console labels days in Pacific Time. Google's guide to using the two together adds that "Clicks and sessions are calculated differently" and that consent opt-outs skew Analytics.
Divide each complete month's total GA4 sessions by that month's Search Console clicks, then learn this seller's normal before hunting for problems. Take the median of those monthly ratios and draw a band 30% either side of it. A recipe blog fed by Pinterest might settle near 2.1 and a software blog living on Google near 1.1; neither figure says anything about honesty. Your questions belong to the months outside the band, and the sharpest ones to months that leave it and snap straight back.
| Likely cause | How the ratio behaves | What the seller can show you |
|---|---|---|
| Cookie consent banner added or tightened | Steps down on launch day and stays lower | The date the banner went live |
| Google tag missing from some templates after a redesign | Steps down; affected pages vanish from GA4 landing pages but keep their search clicks | The redesign date and when the tag was restored |
| Seasonal social peak, such as holiday recipes | Rises for weeks, carried by social referrals rather than search | The same peak a year earlier, on a screen share |
| Newsletter send, giveaway or press mention | Spikes for a few days, then returns | The send date or the article itself |
| Paid placements, traffic exchanges or junk hits sent to the tag | Jumps, often in bursts aimed at one page | Nothing that makes it organic; value the site as if it stops |
Ask your agentFor Briarcliff Recipes, line up total GA4 sessions against Search Console clicks for each complete month of the last 12, find the median ratio, and flag any month more than 30% away from it.
Tools it calls get_ga4_overview get_search_console_summary
The 30% threshold is a buyer's rule of thumb, not a Google standard; loosen it for a site that leans on social. For a second angle, ask for organic search sessions against clicks over the last 90 days. The Organic Arbitrage guide uses a planning range of roughly 0.8 to 1.5 for that ratio and flags anything above 2.0 or below 0.5, while calling its figures illustrative.
Partial months
A 12-month window ending today starts and ends mid-month, and recent Search Console data can lag. Tell your agent to compare full calendar months only, or the last month looks like a collapse.
Worked example: a listing whose traffic does not add up
A buyer reviews a recipe blog listed with a chart of rising sessions. The seller granted GA4 Viewer and restricted Search Console access, and the buyer linked both in DuckDive.
For Briarcliff Recipes, line up total GA4 sessions against Search Console clicks for each complete month of the last 12, find the median ratio, and flag any month more than 30% away from it.
list_websitesfinds the Briarcliff Recipes websiteget_ga4_overviewone session count per day across the trailing year, added up by calendar monthget_search_console_summaryweekly click totals for that year, each week filed under the month it starts in
I left out the partial months at each end, so this covers December through October. Search Console reports clicks by week, so each monthly click figure can be off by a few days. The ratio holds between 1.38 and 1.40 in nine of eleven months. March and July break it: sessions jump while clicks barely move.
| Month | GA4 sessions | Search clicks | Ratio |
|---|---|---|---|
| Dec | 61.2k | 44.1k | 1.39 |
| Jan | 64.8k | 46.9k | 1.38 |
| Feb | 59.3k | 42.7k | 1.39 |
| Mar | 88.4k | 43.5k | 2.03 |
| Apr | 63.1k | 45.6k | 1.38 |
| May | 66.0k | 47.8k | 1.38 |
| Jun | 68.9k | 49.2k | 1.40 |
| Jul | 97.5k | 50.3k | 1.94 |
| Aug | 71.6k | 51.0k | 1.40 |
| Sep | 70.2k | 50.7k | 1.38 |
| Oct | 73.4k | 52.9k | 1.39 |
The median ratio is 1.39, which puts your band at roughly 0.97 to 1.81. Two months land outside it: March at 2.03, 46% above the median, and July at 1.94, 40% above. Each returns to 1.38 or 1.40 the following month. Search clicks, meanwhile, climbed a steady 20% from December to October. That climb is the site's underlying trend; the listing's best months are the two outliers.
I can't split March by source: the channel tool returns totals for a range, not a monthly series. July sits inside the last 6 months, so a referrer that shows up in the 6-month source mix but not the 12-month one is a lead, not proof. Ask the seller for July's source report.
The buyer now holds two dated anomalies to put to the seller in writing. Before you send them, rerun a month or two by hand using our walkthrough for auditing what an agent reports.
Run the ratio on the listing on your desk
Link the property and Search Console site the seller shared, then ask for the month-by-month ratio and the months that broke it.
Free for 2 sites · No credit card
Is the growth earned or bought?
Paid traffic stops when the payments stop, and the seller's budget does not come with the domain. Ask how much of the recent growth still arrives in your first quarter of ownership, with no spend from you.
Start with how GA4 labels a visit. Google's default channel group defines Organic Search as "non-ad links in organic-search results" and Referral as "non-ad links on other sites/apps." That is the gap: a paid placement on a deals site or in a sponsored newsletter, sent without campaign tags, looks like any other link. It lands in Referral, so a channel chart can show bought traffic as free.
Ask your agentWhat share of Briarcliff Recipes' sessions in the last 6 months came from paid or referral sources, and which referrers?
Tools it calls get_ga4_channels
Then check the ad account. The Organic Arbitrage guide advises checking for past Google Ads activity even when the seller says none ran. Ask for two windows, so a recent pause can't hide a year of buying.
Ask your agentThe Larkmoor Pet Supply seller says they switched Google Ads off in the spring. Show me the account's spend, clicks and conversions for the last 90 days and for the last 12 months.
Tools it calls get_google_ads_summary
Last, look at what people searched. Brand-name searches usually follow social posts, press or email, which may leave with the seller. Non-brand searches are the durable kind.
Ask your agentList Briarcliff Recipes' top Search Console queries for the last 90 days and tell me what share of their clicks come from searches containing the brand name versus recipe searches.
Tools it calls get_search_console_keywords
Only the 100 biggest queries come back, and a recipe site's long tail of small searches sits beyond them, so the brand share will read higher than it really is; write it down as an upper bound. Whether a few pages carry most of the traffic is a concentration question; the content site portfolio page covers how to measure it once you own the site.
Worked example: paid traffic presented as organic growth
A pet supply store claims sessions are "up strongly since spring thanks to SEO." The buyer has linked GA4, Search Console and Google Ads read-only access in DuckDive.
Larkmoor Pet Supply credits SEO for its growth since spring. Break the latest 90 days and the stretch before them down by source, and check whether search clicks moved with them.
get_ga4_overviewa six-month daily session series, cut at the 90-day markget_ga4_channelsgrouped by medium, then again by source, for both windowsget_search_console_summaryclicks, cut at that same 90-day markget_google_ads_summarywhat the account spent, and the clicks it bought, in the latest quarter
Sessions rose 63%, but Search Console clicks rose only 3%. The extra sessions arrived as referral and paid search visits, not from Google's organic results.
| Medium | Earlier 90 days | Last 90 days | Change |
|---|---|---|---|
| organic | 17.7k | 18.4k | +4% |
| referral | 1.2k | 16.1k | 13x |
| cpc | 2.1k | 6.3k | +200% |
| all other | 6.6k | 4.2k | -36% |
| Total | 27.6k | 45.0k | +63% |
Organic sessions grew 4% and search clicks went from 21,300 to 21,900, so the two agree: search is flat. A second call, grouped by source instead of medium, names the new referrer: petdeals.example sent 14,900 sessions in the last 90 days and the same 14,900 across the full 6 months, so nothing before that, and it alone is a third of all recent sessions. The Google Ads account spent $4,820 for 5,940 clicks in the last 90 days, which does not fit "switched off in the spring."
The "Earlier" column is whatever remains of the six-month window once the latest 90 days come off, about 92 days, so read it as close rather than exact.
Strip out the referrer and paid search and Larkmoor is a flat store with a healthy organic base. That can still be a fine purchase, priced on the flat number. Ask the seller what the deals placement cost and why the ad account is still spending.
Confirming the property really measures the site for sale
An honest-looking property is worthless if it tracks a sister site or a staging copy. The Organic Arbitrage guide warns that sellers may share a similarly named property, and says to match its ID against the live site's tracking code. Run three checks.
1. The pages agree
The pages with the most Google clicks should also be among the property's top landing pages.
Ask your agentDo Larkmoor Pet Supply's top Search Console pages match its top GA4 landing pages, or does the Analytics property seem to measure a different site?
Tools it calls get_search_console_pages get_ga4_landing_pages
| Page | Search rank | GA4 rank |
|---|---|---|
| / | 1 | 2 |
| /dog-food/grain-free | 2 | 3 |
| /cat-litter/clumping | 3 | 5 |
| /aquarium/filters | 4 | 7 |
| /deals/fall-bundle | not in top 100 | 1 |
Here the property passes: the top search pages all sit in the GA4 top ten. /deals/fall-bundle is the top GA4 landing page with no search presence, which suggests the referral lands there. Ask the seller to confirm.
2. The tag on the live site matches the property
View the live site's page source and find the Google tag ID that starts with "G-". One snag: Google's help page on finding your Google tag ID says you need Editor or higher to see the property's ID in Admin. This is the one job for a live screen share: the seller opens Admin, Data streams, and reads the Measurement ID while you check the source.
3. Search Console sees every part of what you are buying
Google's page on adding a Search Console property explains that a Domain property "Includes all subdomains" and both http and https, while a URL-prefix property includes "only URLs with the specified prefix." If the sale includes a shop. subdomain the seller added last year, or old pages still served on http://, a single https://www. prefix property leaves their clicks out. Part of what you are paying for then goes unchecked, and the sessions-to-clicks ratio sits higher than the site deserves.
Before any comparison, have the seller open Search Console's property menu on a call and read out every property listed for the domain. Match that list against what the sale includes, and ask for a Domain property if one exists. How DuckDive links each property type is on the Search Console integration page.
Connecting one listing, then letting go
DuckDive reads the GA4 property, Search Console site and Google Ads account the seller shares with your Google login, imports their history, and lets your AI agent question it. You get the numbers straight from Google, not an export, and because the history is stored, you can ask new questions a week later without chasing the seller.
-
Send the seller one access request
GA4 Viewer at property level, Search Console restricted user, and Google Ads read-only, all on the Google address you will sign in with. The Google access guide has the exact clicks for the seller.
-
Sign in to DuckDive with that Google account
Use the Google address the seller added you with, or paste the setup prompt from the top of this page into your agent.
-
Create the listing as a website and attach what the seller shared
Pick the seller's property and Search Console site from the list DuckDive shows you, and their Ads account if the listing has ever bought clicks. History lands in DuckDive within minutes in most cases. If you ticked "Also allow actions" while connecting your assistant, it can do the picking for you.
Paste into your agentAdd larkmoorpetsupply.example to DuckDive and link the Analytics property and Search Console site the seller just shared with my Google account. Ask me before linking anything with a different domain in its name.
Tools it calls
add_websitelist_available_sourcesconnect_website_sources -
Run the checks in order
Red flags, the ratio, sources and ad spend, then the property match. Save each answer with its date; they become your written questions to the seller.
-
Close the loop if the deal dies
The seller removes your user in GA4 and Search Console; the access guide shows them where. Then request deletion of the site and everything DuckDive imported for it.
Ask your agentWe walked away from Larkmoor Pet Supply and the seller pulled my Google access. Which of its sources now show as failing, and when did each last import?
Tools it calls
get_sync_status
The rest of the diligence, and where it lives
Everything above settles a single question: are the Google traffic numbers real, and where did they come from. Passing it says nothing about the items below, so give each its own line on your checklist.
| Open question | Why your agent can't answer it here | Where to settle it |
|---|---|---|
| Is the income real? | DuckDive connects to no earnings, affiliate, payment or store accounts | Each income dashboard, live, with the seller signed in, as the Organic Arbitrage guide advises |
| Where are visitors, and on what devices? | Country, device and browser splits sit outside the agent's Analytics tools | GA4's own reports, under the Viewer role you already hold |
| What happened before the trailing year? | Its look-back stops at the trailing year, counted back from today | A call where the seller scrolls GA4 back while you take notes |
| Is there a penalty, or a bought link profile? | Manual actions and backlinks are not part of what DuckDive imports | Search Console's Manual actions report, opened by the seller, plus a backlink tool |
Adding the site to brands you already run? See the holding company page. Buying companies rather than single sites? Start with private equity or the portfolios overview.
Seller pushback and buyer doubts
The seller will only send screenshots and a PDF. Should I walk?
Not on the spot, but treat it as the first red flag. Explain that GA4 Viewer and Search Console restricted access are read-only and can be removed in a minute. A seller with clean numbers rarely refuses after that. If they still say no, price the refusal into your offer or move on.
Could I break something in the seller's accounts by accident?
Not with the roles on your request: GA4 Viewer and a restricted Search Console user can look but not edit. On the DuckDive side there is no delete action and no way to add or remove Google logins, and the few actions that exist only set things up in your own DuckDive account, if you allowed them.
We signed an NDA. Who sees my questions about the listing?
What you type stays between you and your assistant. On DuckDive's side the only trace is a log entry per answered request: when it ran, which website and tool, whether it succeeded, and the request itself with sensitive values masked. The figures it returns are then held by your AI vendor under that vendor's terms, so read those terms against the NDA before you start.
Can my agent see the traffic before the seller started growing it?
Only if that period falls inside the trailing year. The agent works in windows that end today, the widest being 12 months, so an older baseline is out of reach. Daily session counts inside that year still let it isolate the weeks around the listing date. For anything earlier, have the seller walk you through GA4 on a call.
We passed on the listing. How do I hand the data back?
In two steps. The seller removes your Google user, which stops DuckDive importing anything new for that site. Then email hello@duckdive.ai asking for the website and the history already imported to be deleted. If the NDA has a return-or-destroy clause, confirm both steps to the seller in writing.
Does checking one listing cost anything?
A single listing fits the free plan, with no card on file. Weighing several listings side by side means one DuckDive website per listing; the plans below show the price of extra sites.
Do I need a particular AI app for these checks?
If the assistant your deal team uses can plug in a remote MCP server and log in through OAuth, it can run them. Claude does on any plan; ChatGPT does in Business and Enterprise workspaces but not on Free or Plus; Cursor and Claude Code do too. And if nobody wants to wire up an agent for one listing, Drake, built into DuckDive, can field the questions with no connector to configure.
Sources
- Google Analytics Audit for Website Buyers: Due Diligence Protocol for Traffic Verification, Organic Arbitrage (third party), checked November 16, 2026.
- Website Due Diligence Checklist (Free PDF + Excel) + Deal-Specific Framework, Investors Club (third party), checked November 16, 2026.
- Access and data-restriction management, Analytics Help, Google, checked November 16, 2026.
- Managing owners, users, and permissions, Google Search Console Help (the restricted user role), checked November 16, 2026.
- About access levels in your Google Ads account, Google Ads Help, Google, checked November 16, 2026.
- Performance report (Search results): Troubleshooting data discrepancies, Search Console Help, Google, checked November 16, 2026.
- Using Search Console and Google Analytics data for SEO, Google Search Central, checked November 16, 2026.
- Default channel group, Analytics Help, Google, checked November 16, 2026.
- [GA4] Find your Google tag ID, Analytics Help, Google, checked November 16, 2026.
- Add a website or platform property to Search Console, Search Console Help, Google, checked November 16, 2026.
Last reviewed . Numbers in EXAMPLE DATA blocks are fictional.
