Northwind Analytics

Lifetime cap is $5,000.00 but this 30-day window only shows $200.00 spent (4% of the cap in-window).

High — likely a material change in wasted spend, delivery, or efficiency.

Diagnose delivery on cmpn_102 (bids, hints, approved ads, targeting) before raising the lifetime cap.

In the last 30 days this ad recorded $140.00 spend and 90000 impressions, but no clicks. That spend did not produce an engagement.

High — likely a material change in wasted spend, delivery, or efficiency.

Pause ad_506 (or its parent if the whole group is wasted), then fix creative and context hints before reactivating. This audit will not pause it for you.

In the last 30 days this ad recorded $120.00 spend and 15000 impressions, but no clicks. That spend did not produce an engagement.

High — likely a material change in wasted spend, delivery, or efficiency.

Pause ad_510 (or its parent if the whole group is wasted), then fix creative and context hints before reactivating. This audit will not pause it for you.

Daily cap is $40.00 (40000000 micros). Over 30 day(s) the expected spend is about $1,200.00, but Insights show $120.00 (10%).

High — likely a material change in wasted spend, delivery, or efficiency.

On cmpn_104, check bids, context hints, creative review status, and targeting. If delivery is constrained, raise the click bid toward the $3–$5 starting band or add approved ads — do not only raise the daily cap.

Lifetime cap is $8,000.00 but this 30-day window only shows $80.00 spent (1% of the cap in-window).

High — likely a material change in wasted spend, delivery, or efficiency.

Diagnose delivery on cmpn_103 (bids, hints, approved ads, targeting) before raising the lifetime cap.

cmpn_105 is status=active but GET /ad_groups?campaign_id=cmpn_105 returned no groups, so the campaign cannot serve.

High — likely a material change in wasted spend, delivery, or efficiency.

Create at least one ad group under cmpn_105, or pause the campaign if it should not spend.

Context hints on the ad group tell ChatGPT when the ad may be useful. They are not exact-match keywords, but an empty list gives the auction no placement guidance from you.

High — likely a material change in wasted spend, delivery, or efficiency.

Set context_hints on adgrp_302 to 3–8 specific descriptions of prompts or jobs-to-be-done (short sentences, not single generic words).

max_bid_micros is 1200000 ($1.20 CPC). OpenAI’s Help Center starting range for CPC is $3.00–$5.00. A bid under $3 often under-competes in this auction.

High — likely a material change in wasted spend, delivery, or efficiency.

On ad group adgrp_302, raise bidding_config.max_bid_micros from 1200000 ($1.20) to at least 3000000 ($3.00), then watch actual CPC.

max_bid_micros is 2500000 ($2.50 CPC). OpenAI’s Help Center starting range for CPC is $3.00–$5.00. A bid under $3 often under-competes in this auction.

High — likely a material change in wasted spend, delivery, or efficiency.

On ad group adgrp_306, raise bidding_config.max_bid_micros from 2500000 ($2.50) to at least 3000000 ($3.00), then watch actual CPC.

Ad groups with a single (or zero) approved variant cannot test titles, bodies, or images against each other, so weak creative keeps all of the spend.

High — likely a material change in wasted spend, delivery, or efficiency.

Get at least one ad in adgrp_303 to review_status=approved (edit rejected ads so they can be re-reviewed), and add a second distinct variant.

review_status=rejected means the ad violated ads policy and will not serve until it is edited and re-reviewed.

High — likely a material change in wasted spend, delivery, or efficiency.

Edit the creative on ad_505 to comply with ads policies and resubmit. Blobbo will not edit the ad.

In the last 30 days this ad recorded $60.00 spend and 40000 impressions, but no clicks. That spend did not produce an engagement.

Medium — a meaningful efficiency, coverage, or structure improvement.

Pause ad_507 (or its parent if the whole group is wasted), then fix creative and context hints before reactivating. This audit will not pause it for you.

All 2 hint(s) are under 3 words / 18 characters (“saas”, “tools”). Official docs allow descriptions or keywords; single generic tokens usually under-specify when the ad is useful.

Medium — a meaningful efficiency, coverage, or structure improvement.

Replace the hints on adgrp_303 with fuller descriptions, e.g. turn “saas” into a sentence about the situation or prompt in which this product should appear.

All 1 hint(s) are under 3 words / 18 characters (“webinar”). Official docs allow descriptions or keywords; single generic tokens usually under-specify when the ad is useful.

Medium — a meaningful efficiency, coverage, or structure improvement.

Replace the hints on adgrp_306 with fuller descriptions, e.g. turn “webinar” into a sentence about the situation or prompt in which this product should appear.

These hints appear in more than one group: “saas”. Overlapping hints make groups compete for the same prompts instead of covering distinct situations.

Medium — a meaningful efficiency, coverage, or structure improvement.

Give each overlapping group in adgrp_303, adgrp_304 a distinct hint set. Keep shared themes only where the offer is truly identical, otherwise split or merge the groups.

Ad groups with a single (or zero) approved variant cannot test titles, bodies, or images against each other, so weak creative keeps all of the spend.

Medium — a meaningful efficiency, coverage, or structure improvement.

Add at least one more distinct chat_card in adgrp_302 (different title/body or image) so you have 2–4 approved variants.

Ads ad_506, ad_507 share the same title and body, so they are not useful variants.

Medium — a meaningful efficiency, coverage, or structure improvement.

Keep one of ad_506 / ad_507 and change the title, body, or image on the others — or pause the duplicates.

Ad groups with a single (or zero) approved variant cannot test titles, bodies, or images against each other, so weak creative keeps all of the spend.

Medium — a meaningful efficiency, coverage, or structure improvement.

Add at least one more distinct chat_card in adgrp_305 (different title/body or image) so you have 2–4 approved variants.

Ad groups with a single (or zero) approved variant cannot test titles, bodies, or images against each other, so weak creative keeps all of the spend.

Medium — a meaningful efficiency, coverage, or structure improvement.

Add at least one more distinct chat_card in adgrp_306 (different title/body or image) so you have 2–4 approved variants.

The API does not expose a last-served timestamp. This check uses updated_at/created_at plus weak or thin performance in the audit window as a stale-creative signal.

Medium — a meaningful efficiency, coverage, or structure improvement.

Create a new variant alongside ad_504 (new title, body, or image) and keep the original paused if the new one wins.

Title “Try it” is 6 characters (API allow 3–50). Short titles often under-explain the offer in a chat card.

Low — incremental hygiene; useful, but smaller than the items above.

Rewrite the title on ad_504 to a clearer 15–50 character offer.

Body is 19 characters (API max 100). That leaves little room to explain why someone should click.

Low — incremental hygiene; useful, but smaller than the items above.

Expand the body on ad_504 toward 40–100 characters with a concrete benefit.