You do not need to commit to anything to find out whether you are bleeding spend. A read-only week is enough. This is the diagnostic loop we run with every new operator before recommending automation, and it works whether or not you ultimately use Campaign-IQ to execute on it.
Day 1 — connect read-only
Connect Shopify (or your equivalent storefront) and your ad platforms with read scopes only. No spend access, no write permissions, no bid changes. The point of week one is observation.
Days 2–3 — rank in POAS
Rank every campaign, ad set, and SKU by POAS over the last 30 days. The bottom quintile is where the bleed lives. Almost every operator finds at least one campaign that has been losing money for months and was hiding under healthy aggregate ROAS.
Days 4–5 — segment by reason
A losing POAS is a symptom; the reason matters. Split the bottom quintile into four buckets — wasted spend has different fixes by category.
- Margin drag — the campaign drives volume on a low-margin or heavily discounted SKU.
- Invalid traffic — high CTR, near-zero add-to-cart, concentrated in a few placements.
- Non-incremental — branded search and returner audiences taking credit for organic checkouts.
- Genuine under-performance — the placement just does not work for your offer.
Days 6–7 — size and decide
Sum the spend in each bucket. Most Shopify operators find 12–22% of last month’s ad spend is sitting in one of the four buckets. That is your savings ceiling — what you would recover if you eliminated the bleed entirely. The realistic recovery in the first quarter is usually 60–80% of that, because some of the spend has hidden incrementality you only see when you turn it off.
Use the size to decide. If the bleed is meaningful, automation pays for itself in week two. If it is small, you have just earned the confidence that your stack is already disciplined.
- You can size wasted spend in a week without changing any campaign.
- POAS-ranking the bottom quintile and splitting by reason tells you what to fix, not just how much.
- The audit is useful whether or not you adopt automation — it is a read on whether the bleed exists.