Agency CTR Analysis

1. The headline Sales is reacting to

Blended CTR (every visit and click counted together) does dip in June:

Taken at face value this reads as "more people land on AD pages but fewer click through to agencies." That reading is wrong.

2. Why it's misleading: clicks held up, impressions spiked

CTR is clicks ÷ impressions. Split the two apart and the cause is obvious – the denominator moved, not the numerator.

• Clicks to agencies held up – ~13,000–17,000/week, and actually rose to ~20,000 in the week of 8 June. People are not clicking less.
• Impressions spiked ~2.5× – from ~174k to 441k in the week of 8 June.
• So CTR fell from ~8% to ~4.6% purely because impressions outran clicks.

The 8 June impression spike is Direct-desktop, single-page, non-clicking, and abrupt – the fingerprint of bot / low-intent traffic, not a change in real demand or in the page. The worst single sessions are plainly automated: one fired thousands of impressions across hundreds of AD pages. None of this lines up with the tabs (changed in March and on 22 May).

Deliverables 1–5 below use clean human traffic – spam removed (Direct sessions with ≥50 pageviews, plus crawlers that load ≥10 distinct AD pages in a session) – over 27 Apr – 7 Jun, before the 8 June impression spike. The same filter is applied to every period, so the comparisons are fair.

Deliverable 1 – Drop-off: where do users go from AD pages?

Outcomes for human sessions that saw an AD listing (one row per session, by its furthest action):

• ~23% engage an agency (18.5% click through to a website, 4.4% open a profile, 0.1% submit a contact form).
• Of those who don't, almost all simply leave: 69% exit with no agency action, and only 8% browse deeper into AD (filters / other categories). People aren't wandering off to other parts of AD – they just leave.

Deliverable 2 – Do the agency-card tabs hurt CTR?

The 22 May change added tabs to sponsor cards. If tabs hurt, sponsor CTR would have fallen across that date. It didn't:

• Sponsor-card CTR held steady at ~9–10% right across 22 May – no decline. Sponsor cards (which have the green "Visit website" CTA) convert far better than non-sponsor cards (~2–5%, profile-link only).
• Verdict: tabs are not hurting agency CTR. The overall "drop" is the impression spike, not the tabs.

A supporting signal (associative, not causal): sessions that open at least one tab reach an agency 57.8% of the time, versus 20.5% for sessions that don't – nearly 3× higher. Tab use goes hand-in-hand with higher intent; it certainly isn't dragging conversion down.

Deliverable 3 – How users use the tabs, and the in-card contact form

First, the big picture – how many visitors open a tab at all:

Only about 1 in 20 AD-listing sessions opens any tab (2,535 of 49,821 sessions, 27 Apr – 7 Jun). The tabs matter to a focused minority; for everyone else the card's headline, rating and CTA carry the click.

Is it worth keeping the tabs? The open question is whether the tabs earn their place, given they add to the DOM and can weigh on Core Web Vitals. The data speaks to the benefit side of that trade-off:

• Reach is small – only ~5.1% of AD-listing sessions open any tab.
• But that minority is high-intent – tab-opening sessions reach an agency 57.8% of the time versus 20.5% otherwise (associative, not proof the tabs cause it), and the Contact tab alone captures ~85–130 in-card leads/week.
• Tabs aren't dragging CTR down – sponsor CTR held steady across the 22 May tab addition.

The data suggests the tabs aren't hurting conversion and do serve a small, high-intent slice of users. What this analysis can't see is the other side of the scale – the Core Web Vitals / DOM cost. A clean decision would need that performance cost quantified, and ideally a test of whether removing the tabs changes conversion for the ~5% who use them.

Among the sessions that do use tabs, here's where they go:

• Services dominates (~a quarter of all tab clicks), then Clients, Overview, Contact and Portfolio in a tight band; Reviews is least used.

Does the in-card contact form divert clicks away from agencies?

• The Contact tab draws ~300–660 clicks/week and the in-card contact form captures ~85–130 submits/week – versus thousands of agency clicks/week. The in-card form is a small share of volume.
• A contact-form submit is a captured lead, not a lost agency click. There is no sign the contact form is meaningfully eating into clicks to agencies.

Deliverable 4 – 'View profile' vs green 'Visit website' on sponsor cards

On sponsor cards, how clicks split between the two buttons (27 Apr – 7 Jun, human traffic):

• View profile is heavily used: ~12,000 clicks, about 1 in 5 (21%) of sponsor-card clicks, against ~45,000 for the green "Visit website" button. (A further ~5,000 view-profile clicks happen on non-sponsor cards, where it's the only CTA.)
• Recommendation: do not remove the 'view profile' button on sponsor cards. It carries a meaningful share of engagement; removing it would cut roughly one in five sponsor-card clicks.

Deliverable 5 – Whole-AD vs the 24 priority "moneymaker" pages

• The 24 moneymaker pages convert far better than the rest of AD and are rising – CTR ~12% climbing to ~22%, versus ~6–7% for the rest. No drop on the priority pages.

Recommendations

1. Review the current AD report to exclude spam and bots inflating agency-card impressions. Data & Automation team's action item.
2. On keeping the tabs: the data suggests they aren't hurting conversion – sponsor CTR held steady across the 22 May addition and tab use correlates with higher intent – but this analysis doesn't capture their Core Web Vitals / DOM cost. Weigh that measured engagement value against the performance cost before deciding.
3. Keep the 'view profile' button on sponsor cards. It draws ~1 in 5 sponsor-card clicks, so it's actively used rather than ignored.
4. If the tabs are kept, keep the in-card contact form too – it captures leads without measurably reducing agency clicks.

Notes

• Spam excluded from all figures: Direct sessions with ≥50 pageviews, plus sessions loading ≥10 distinct AD pages (crawlers). Notable spam days: 14 May, 19 May; the larger impression spike runs from 8 June.
• Data: GA4 → BigQuery, 27 Apr – 17 Jun 2026. Deliverables 2/4/5 use 27 Apr – 7 Jun (before the 8 June spike).

Last refreshed 17 seconds ago