Internal · Numbers through 18 Sep · Status 21 Sep
Before paid ads, the site was smaller and deeper. Reddit bought a lot of sessions. Visits got shallower. Contact us did not move. Filtered trials went from 31 in July to 42 in August, a modest improvement that will continue improving with time.
Google Search is the intent test. When Search was on and Reddit volume fell, visits looked a bit deeper. That is a reason to keep testing, not a conversion win.
Website
Two pictures of the same site. Left is the true no-ads window. Right is the first weeks after Reddit went live. Unique users are period uniques. Clarity counts some bot traffic. The direction is the same either way.
14 Jul-4 Aug · no paid traffic
5-26 Aug · Reddit live from 5 Aug
More people, shallower visits, same Contact us. Reddit bought volume. It did not buy depth in this window.
Tests
Three channels, three jobs. Google is the intent test. Reddit is volume and geography. OpenAI is still an open question. Spend below is through 18 Sep, as billed.
The intent test
US ConnectWise and Dynamics 365 Search. Snapshot 3-17 Sep: R12,672 and 98 clicks. New landing pages are live. Taylor L is restarting Search.
Search is where someone types a job (automation, API, integration), not where they scroll a feed. That is why we keep testing it even though clicks are a fraction of Reddit.
Brand Search and Display are being scoped. They are not results in this snapshot.
Volume and geo
The first pilot optimized for cheap ZA clicks (about 98% of impressions). The small US slice of the same pilot had about 5x the CTR and a much higher CPM.
A US-only Webinar Booster is the first attempt to stop that cheap-ZA default. Next: US vs ZA, homepage vs landing page, then tighter pages. Creative that got clicks: AI meme, Drake ERP/CRM, Ira Green quote, icon plus meme text.
Open question
Campaign A and Campaign B. $1,026, 0.66% CTR, about $5 CPC. chatgpt.com referrer still around 10. Visitors who do arrive tend to stay. Running again at a modest daily cap. Not a site-lift story yet.
When Reddit spend paused and Google was on (19 Aug-2 Sep vs 3-17 Sep), sessions fell about 25% and visits looked a bit more useful. Read this as an early intent signal, mixed with geography. Reddit traffic is mostly ZA. Google is US Search. We cannot yet separate those two.
Direction of travel
This is a snapshot, not a locked media plan. Mix, geos, and pages may still change.
Full tables from the 18 Sep cut. The main story above is the briefing. This is the backup.
True no-ads baseline vs Reddit launch. Unique users are period uniques.
| Metric | 14 Jul-4 Aug | 5-26 Aug | Read |
|---|---|---|---|
| Sessions | 2,703 | 7,051 | Volume up |
| Unique users | 1,783 | 6,046 | Volume up |
| reddit.com referrer | 0 | 2,141 | Matches Reddit go-live |
| Pages per session | 1.63 | 1.24 | Shallower |
| Scroll depth | 43.7% | 23.1% | Shallower |
| Contact us events | 44 | 46 | Flat |
| Bot sessions | 1,355 | 2,497 | Rose with volume |
| Free-trial funnel | 0 | 0 | Signup finishes on the app |
Matched 15-day windows around Google go-live. Bot sessions in parentheses under sessions.
| Metric | 19 Aug-2 Sep | 3-17 Sep | Read |
|---|---|---|---|
| Sessions | 4,562 (1,141 bot) | 3,408 (774 bot) | Volume down about 25% |
| Unique users | 3,948 | 2,717 | Down with volume |
| New users | 84.8% (3,868) | 77.3% (2,633) | Less new-heavy |
| Returning users | 15.2% (694) | 22.7% (775) | More repeats |
| Pages per session | 1.30 | 1.41 | Deeper |
| Scroll depth | 25.0% | 30.4% | Deeper |
| Active time on site | 37 sec | 51 sec | Up |
| Bot sessions | 1,141 | 774 | Improved with lower volume |
| Contact us | 33 | 47 | Modest |
| Log-in | 81 | 109 | Up |
Start Building Button Click and GAds Form Submit both moved 3 to 56 in lockstep. That is one Clarity click event, counted twice. It likely includes internal page testing. Do not read it as 56 leads. Daniel said on 17 Sep that the landing-page form itself was not completing. Google showed 0 in-platform conversions on 98 clicks.
| Event | 19 Aug-2 Sep | 3-17 Sep | Read |
|---|---|---|---|
| Start Building Button Click | 3 | 56 | Click event. Not a conversion. |
| GAds Form Submit | 3 | 56 | Same jump. Same caveat. |
| Referrer | First window | Second window | Read |
|---|---|---|---|
| www.reddit.com | 1,365 | 482 | Reddit budget pause plus Google live |
| www.google.com | 327 | 460 | Up with Google Ads |
| www.flowgear.net | 433 | 558 | Internal / self-referral |
| www.bing.com | 36 | 59 | Up |
| chatgpt.com | 10 | 9 | Flat. No OpenAI lift yet. |
Real spend starts 3 Sep. In this cut both US Search campaigns were paused while new landing pages shipped, not for performance. As of 21 Sep those pages are live and Search is restarting. In-platform conversions: 0 in this snapshot.
| Campaign | Status in cut | Cost | Clicks | CTR |
|---|---|---|---|---|
| US | ConnectWise | Search | Paused in cut | R6,879.68 | 49 | 3.23% |
| US | Dynamics 365 | Search | Paused in cut | R5,792.15 | 49 | 2.76% |
| Ad group | Campaign | Cost | Clicks |
|---|---|---|---|
| ConnectWise Automation | US | ConnectWise | Search | R4,440.50 | 31 |
| Dynamics 365 Automation | US | Dynamics 365 | Search | R2,905.44 | 23 |
| ConnectWise API | US | ConnectWise | Search | R1,562.16 | 12 |
| Dynamics 365 API | US | Dynamics 365 | Search | R1,534.94 | 12 |
| Dynamics 365 Integration | US | Dynamics 365 | Search | R1,351.77 | 14 |
| ConnectWise Integration | US | ConnectWise | Search | R877.02 | 6 |
| Keyword | Match | Cost | Clicks | CTR |
|---|---|---|---|---|
| connectwise automation software | Phrase | R2,826.05 | 18 | 3.13% |
| dynamics 365 automation software | Phrase | R2,152.59 | 17 | 2.17% |
| connectwise api | Phrase | R1,466.46 | 11 | 6.04% |
| connectwise automation | Phrase | R1,100.22 | 9 | 2.60% |
| dynamics 365 api integration | Phrase | R993.00 | 7 | 3.85% |
| microsoft dynamics 365 integration | Phrase | R574.85 | 6 | 5.88% |
Auction in this cut: Flowgear impression share 17.5% vs connectwise.com 21.8%. Desktop about 67% of spend. Age 35-54 dominates. Peak days Thu/Fri.
Reddit went live 5 Aug. Spend paused about 31 Aug-6 Sep after hitting budget while focus shifted to Google, then came back. That pause sits in the later website window and helps explain the reddit.com referrer drop.
| Campaign | Country | Impressions | Clicks | CTR | CPC | Spend |
|---|---|---|---|---|---|---|
| Flowgear Aug 2026 Pilot | ZA | 1,696,780 | 12,693 | 0.75% | $0.12 | $1,503.96 |
| Flowgear Aug 2026 Pilot | US | 3,337 | 128 | 3.84% | $0.19 | $24.87 |
| Flowgear 16 Sep Webinar Booster | US | 10,168 | 102 | 1.00% | $1.03 | $105.10 |
ZA carries about 98% of volume at low CPC. US slice of the same pilot has about 5x the CTR. Webinar Booster (US, from about 16 Sep) is early in this window.
| Campaign | Spend | Impressions | Clicks | CTR | Avg CPC |
|---|---|---|---|---|---|
| Campaign A | $971.94 | 29,917 | 196 | 0.66% | $4.96 |
| Campaign B | $54.37 | 1,342 | 19 | 1.42% | $2.86 |
Both share ad group name Integration Platform Evaluation (1 Aug-18 Sep). Conversions 0 in-platform. Clarity chatgpt.com referrer still single digits.
GA4 channel mix is not in this cut.