The Accounts, Not The Mockups

Search at a $3,000 daily budget

Spend

$63.9K

Clicks

36,380

Impressions

597K

Keyword-level view of a US search build. Four head terms carry the account, each reported on its own cost and click volume rather than averaged into a campaign total — which is the only way to see which term is actually paying for the others.

Twenty-three campaigns, with the totals row

Campaigns

23

Total spent

$449.26

Reach

60,803

The whole table this time rather than a slice of it, down to the results row: 23 campaigns split by geo and budget tier, $449.26 spent, 60,803 people reached, 965 link clicks. Per-campaign frequency sits between 1.00 and 1.34 while the blended figure reads 1.46 — the gap between those two numbers is the argument for splitting by market, because the blend is the only one a summary report would have carried. Cost per result reads as a dash across the set: attribution had not landed at the time of capture, and it is published with that gap showing rather than cropped away. The account label is blacked out.

Multi-geo push campaign set

Impressions

535K

Clicks

42K

Bid range

$0.01–$0.15

Country-split campaigns with the bid set per geo rather than globally — $0.01 where the inventory is cheap, $0.15 where it converts. At this end of the market, bid discipline is most of the margin.

Format spread across one offer

Campaigns

15

Formats

4

Bid models

CPA · CPM · SmartCPC

Fifteen campaigns queued across OnClick, Push, In-Page Push and Calendar placements, each on its own bid model — CPA, CPM, SmartCPC — so no single format ends up setting the price for the rest.

Micro-bidding by subscriber list

Lists managed

12+

Best winrate

88.08%

eCPC

$0.008–$0.009

Per-list bidding on a broadband offer. Lists that convert get a custom bid, the rest get paused. Winrate and eCPA are the only two columns that decide it — everything else is noise at this granularity.

Twenty keywords, read one at a time

Spend

$24.4K

Clicks

5.5K

Keywords

20

The whole keyword table from a US, CA and AU build, scrolled end to end and rejoined into a single view. Paused terms are left sitting next to eligible ones on purpose — a keyword that stopped serving still spent, and dropping it flatters the average. Account and campaign labels are stripped out of the left column.

Which campaign each keyword sits in

Terms shown

8 of 20

Geo split

US · CA

Top term

604 clicks

The same account as the table above, read from the app instead — where every keyword carries its campaign and geo on the line underneath it. Two adjacent screens rejoined side by side, top eight terms by impressions. Half are eligible and half sit under a paused campaign, which is the reason to read it this way: the paused CA terms still spent, and only the per-keyword label says which geo that money went to.

A campaign before it spends anything

Date range

Mar 11 – Apr 6

Ad groups

1

Served

Not yet

Keyword set and ad copy staged against a column of $0.00. Everything reads zero because nothing has served yet, and that is the state a build is worth reviewing in — copy and landing promise can still be changed for free. The display URL is redacted.

Opening budget held at $25

Budget

$25 total

Bidding

Maximize clicks

Optimization

79.3%

A search campaign in bid-strategy learning on a $25 total budget. The optimization score sits at 79.3% and stays there — closing the last twenty points means taking Google’s recommendations wholesale, and most of them widen targeting rather than sharpen it.

Volume first, then the ad copy

Top term

165K / mo

Terms sized

11

Output

Ad copy set

Eleven crypto terms sized by monthly volume before a single headline gets written. “Crypto wallets” at 165K carries a different intent to “best crypto to buy now” at 60.5K, and one ad cannot serve both — so the copy set is drafted per intent tier, not per campaign.

The filter deciding, click by click

Clicks in flight

113

Carrier traffic

→ Offer

VPN / proxy

→ White

One flight, 113 clicks, with the two columns that actually settle the argument left open: FILTER and PAGE. Consumer mobile carriers — T-Mobile, Verizon, Comcast, AT&T — resolve to Offer. The rows flagged VPN/Proxy, and the Google datacentre ranges beneath them, resolve to White. This is what a working setup looks like from the inside: not a claim about compliance, a per-click decision that can be audited after the fact. The visitor IP column is blacked out — those are real people’s addresses.

Three campaigns, named for their jobs

Campaigns

3

Flows

1 each

Served

Not yet

The architecture stated in the campaign names rather than buried in a rule set: a bot bucket, a safe page flow and an offer page campaign, one flow each, two of them bound to Google Ads as the source. Captured at zero clicks, because a build is worth reviewing before it serves — that is the only point at which changing it costs nothing. The tracker admin URL is cropped out.

Server-to-server postback, wired all the way

The network side of the tracker: one offer parameter carrying the click ID out, and a postback URL bringing status, payout, currency and all five sub-IDs back. Leave sub2 through sub5 unmapped and a conversion still lands — it just lands with no idea which placement earned it, which is how a source gets cut for looking unprofitable. The tracker host and campaign token on the first line are blacked out.

Click-level logging: TikTok into Brazil

Clicks logged

128

Source

TikTok

Geo depth

State + city

Every click stored with source, state, city, OS, browser and connection type. This is what makes traffic quality arguable after the fact instead of guessed at — and what a filtering rule gets written against.

Tracker domains on HTTPS-only

Campaign domain attached to the tracker with HTTPS-only enforced, mid DNS propagation. The server address is redacted — publishing a live tracker IP is an invitation, not a credential.

Nameserver delegation

NS records delegating the campaign domain to DigitalOcean so records move on my schedule instead of the registrar’s — with MX preserved first, because pointing nameservers before mail records is how a domain loses its email.

Conversion action before first click

The conversion action exists, the page-view tag is not yet verified, and the ad is enabled at zero impressions. Order matters here — a campaign that starts serving before its tag reports has already bought data it cannot read.

One profile, one static ISP proxy

Profiles

1 / 12

Static ISP

1

Datacenter

0

Profile isolation in an antidetect browser with a single static ISP proxy bound to it, the datacenter and imported pools deliberately empty — an exit address that changes every session is a fingerprint of its own. The proxy address and the account label are blacked out.

Keitaro Expert, licensed and current

Edition

Expert

Expires

Aug 4, 2026

The tracker behind the click logs, running on a paid Expert licence rather than a shared or cracked install. The key and account email are blacked out — a licence key is a credential, and what this page is here to show is the edition and the expiry.

A licence per instance, not one shared install

Licences

2

Edition

Starter

Paid to

Sep 20, 2026

Two Starter licences held alongside the Expert instance above — one paid and running to September 20, 2026, one unpaid and deliberately idle. Separate licences per instance rather than one install shared across engagements is the reason a client’s click data never sits in another client’s tracker. Both keys are blacked out; a licence key is a credential, and what this page is here to show is the edition and the expiry.

Grooming storefront, front to footer

A full storefront build in one scroll: hero, category split across devices and grooming essentials, two product grids, a trust row and email capture. Every block earns its place on the way to the cart.

Catalogue-wide product SEO

Score

97 / 100

Coverage

Full catalogue

Depth

3 pages

Focus keyword, title and meta written per product and scored across the catalogue rather than on the hero SKUs alone — three pages deep, every row sitting at 97.

Search Console verified at DNS level

Property verified through the domain name provider rather than by file upload, so verification survives a theme change, a replatform, or anyone clearing the storefront root.

988K clicks across three months

Clicks

988K

Impressions

6.21M

Avg. CTR

15.9%

Three months of organic search on one property. The interesting pair is the last two tiles: holding a 15.9% average CTR from average position 8.3 means the titles and descriptions are earning clicks the ranking on its own would not.

Ten years of product-page traffic

Clicks

1.25M

Impressions

192.73M

Purchases

876

Ads and organic plotted on one axis across the whole history of a catalogue: 1.25M clicks against 192.73M impressions. The orange spikes are campaign windows and the flat stretches are the store running on organic alone — which is the comparison this report exists to make, and the reason paid work here gets scoped against an existing baseline rather than presented as a launch.

$9,130 off 285 sessions

Total sales

$9,130

Orders

13

Sessions

285

A seven-day window with the week before drawn in behind it as a dotted line. 285 sessions and 13 orders puts average order value near $700, which is what makes a 1.4% conversion rate the right number to leave alone — and the two backlog tiles underneath, 50+ orders to fulfil and 50+ payments to capture, say the week was not a one-off spike.

A client reporting $43.6K back

Total sales

$43,600.90

Sessions

4,140

Live visitors

20

The client’s own Shopify dashboard, sent over as it happened — 4,140 sessions against $43,600.90 in total sales, twenty people on the store at that moment. Their name and handle are cropped out; the numbers and the message are theirs, unedited.