CAMPAIGN PERFORMANCE ANALYSIS - SYNTHETIC SAMPLE DATA ===================================================== These three files accompany the guide at https://www.findanomaly.ai/campaign-performance-analysis They are SYNTHETIC SAMPLE DATA generated by a seeded script (seed 20261008). They contain no real ad account, analytics property or customer data. Every row carries data_label = "synthetic sample data". You may reuse them freely. Scenario: a fictional six-person online store selling outdoor gear. Eight Google Ads campaigns, two email programs and two paid social campaigns, 91 days from Monday 2026-06-29 to Sunday 2026-09-27. The export was pulled at about 10:00 on Sunday 2026-09-27, the day before a budget meeting. Target: return on ad spend (ROAS) of 3.0, i.e. $3 of conversion value per $1 of ad spend. FILES ----- 1. campaign-google-ads-daily.csv (4,095 rows) One row per date x campaign x network x device, shaped like a Google Ads campaign report. date, campaign_id, campaign, campaign_type, network, device, impressions, clicks, cost, conversions, conv_value -> dated by the day of the ad click (the Google Ads default) conversions_by_conv_time, conv_value_by_conv_time -> dated by the day the conversion happened data_label 2. campaign-ga4-session-campaign-daily.csv (1,274 rows) One row per date x session campaign, shaped like a GA4 traffic-acquisition export. date, session_campaign, session_campaign_id (Google Ads campaigns only), session_source_medium, sessions, engaged_sessions, key_events_purchase, purchase_revenue, data_label GA4 dates a purchase on the day it happened and credits it under its own attribution model. Includes email, paid social, (organic) and (direct) rows for context. 3. campaign-other-channels-export.csv (364 rows) Email and paid social numbers as each platform reports them. date, channel, campaign, spend, clicks, platform_conversions, platform_revenue, data_label Email spend is an allocated share of the email tool's monthly fee. PLANTED EFFECTS (what the guide teaches) ---------------------------------------- 1. Conversion lag of 0-7 days. Google Ads books a conversion on the click date; GA4 and the *_by_conv_time columns book it on the day it happened. 2. A renamed campaign. campaign_id 2007 is "Search - Hiking Boots" until 2026-08-16 and "Search - Footwear" from 2026-08-17, in both files. Join on campaign_id, not the name. 3. A partial last day. 2026-09-27 holds about 40% of a normal day's clicks. 4. A noisy high-CPA campaign. "Search - Trail Running" has 3 conversions in its last complete week (cost per conversion $85.75) but a 12-week ROAS of 3.28. Its last four weeks hold only 20 conversions. 5. One real outlier. "Search - Tents" on the Search partners network from 2026-08-24: spend rises from about $108 to about $612 a week and the conversion rate falls from 2.13% to 0.33%. ANSWER KEY (complete weeks 1-12 = 2026-06-29 to 2026-09-20 unless stated) ------------------------------------------------------------------------- Google Ads totals: cost $84,542.09; clicks 101,265; conversions 2,710 (click date), 2,722 (conversion time); conversion value $315,669.44 (click date), $316,950.77 (conversion time). GA4, same campaigns: 92,121 sessions (9.0% fewer than clicks); 2,283 purchases; $265,567.35 revenue. GA4 purchases vs Ads conversions: -15.8% (click date), -16.1% (conversion time). Dating moves 276 conversions (10.2%) between campaign-weeks but barely changes the 12-week total. Week 13 (incomplete): Ads click-date conversions 136; Ads conversion-time 187; GA4 156. Dashboard ROAS (click date): week 12 = 3.40, week 13 = 2.48, weeks 1-12 = 3.73. Week 13 ROAS by conversion time = 3.45. ROAS by conversion time: weeks 1-8 = 3.94; weeks 9-12 = 3.38; weeks 9-12 without Search - Tents on Search partners = 3.65. Per campaign, 12 weeks, conversion time: campaign cost conv value CPA ROAS GA4 purchases GA4 gap Display - Remarketing 9,690.23 170 19,192.61 57.00 1.98 90 -47.1% Search - Backpacks 12,037.90 425 43,298.95 28.32 3.60 364 -14.4% Search - Brand 4,383.54 429 50,353.58 10.22 11.49 398 -7.2% Search - Footwear 8,829.11 217 32,868.03 40.69 3.72 184 -15.2% Search - Tents 18,200.63 271 58,440.62 67.16 3.21 232 -14.4% Search - Trail Running 3,069.38 70 10,080.60 43.85 3.28 64 -8.6% Shopping - All Products 21,114.92 672 77,291.95 31.42 3.66 558 -17.0% Shopping - Clearance 7,216.38 468 25,424.43 15.42 3.52 393 -16.0% Last four complete weeks (9-12), conversion time: ROAS / conversions Display 1.78/46; Backpacks 3.56/147; Brand 11.61/136; Footwear 3.46/71; Tents 2.31/79; Trail Running 2.48/20; Shopping - All Products 3.41/197; Shopping - Clearance 3.76/158. Search - Tents by network, conversion time: weeks 1-8 Google search ROAS 3.87 (conv rate 2.74%); Search partners ROAS 3.21 (2.13%) weeks 9-12 Google search ROAS 3.21 (2.28%); Search partners ROAS 0.52 (0.33%), cost $2,449.79 Decision rule used in the guide (target ROAS 3.0, conversion-time columns, complete weeks only): Scale = ROAS 3.5 or more in both the 12-week and last-4-week windows Cut = ROAS below 2.5 in both windows Fix = below 2.5 in the last 4 weeks only (find the segment that changed) Hold = everything else, or fewer than 30 conversions in the last 4 weeks Result: Scale - Backpacks, Brand (capped by brand searches), Clearance; Hold - Footwear, Shopping - All Products, Trail Running (too few conversions); Fix - Tents; Cut - Display - Remarketing. Other channels, 12 weeks: platforms report 3,759 conversions in total (Google Ads 2,710 + email 443 + paid social 606); GA4 credits 2,963 purchases to the same channels: 796 fewer, or 21.2% below the platforms' total (the platforms claim 26.9% more than GA4). Every figure above can be recomputed from the three CSVs with a spreadsheet pivot table: filter week 1-12 (date 2026-06-29 to 2026-09-20), group by campaign_id, and sum the columns.