The problem
The brand sold shoes online and advertised heavily on Facebook and Instagram. The account had grown to more than 80 fragmented campaigns. Audiences overlapped, budget was spread thin, and campaigns kept dropping back into Meta's learning phase. The objective was to improve results without increasing the monthly budget.
More than 80 live campaigns, many targeting the same people with different ads.
Constant learning-phase resets from edits to small, fragmented campaigns.
Budget split into pieces too small for any campaign to exit learning.
No clear line between prospecting, retargeting and past customers.
With so many campaigns, it was hard to see which ones actually earned their spend.
What we built
We treated the account as a system to be simplified rather than a budget to be increased. Every campaign was audited and mapped to the stage of the funnel it served. Then the account was rebuilt as four funnels: prospecting, lookalikes, retargeting, and high-intent audiences such as carts and past buyers. Overlapping audiences were separated and budgets were consolidated into the campaigns with the strongest purchase signals.
Account structure
Same ten-day window, one month apart
Before · 80+ campaigns
- Overlapping audiences bidding against each other
- Frequent learning-phase resets
- Budget split into pieces too small to work
- Prospecting and retargeting mixed together
- Results, 10 days
- 373
- Objectives
- Mixed
After · 4 funnels
- Prospecting
- Lookalikes of buyers
- Retargeting: visitors and engagers
- High intent: carts and past buyers
- Purchases, 10 days
- 643
- Cost per purchase
- $4.50
Account audit
Every campaign classified by audience, objective and the purchase signals it had produced.
Four-funnel structure
Prospecting, lookalikes, retargeting and high intent, each with its own budget and measurement.
Audience separation
Website visitors, engaged followers, add-to-cart users and previous customers split so they are never bid on twice.
Budget consolidation
Spend moved from dozens of small campaigns into the few with stable purchase signals.
Signal-led budgets
Spend concentrated where purchase signals were strongest, so campaigns could leave the learning phase.
Continuous monitoring
Performance compared by funnel stage, with budget shifted toward the structures producing purchases.
How it works
From trigger to result.
How the system works
7 stages
01Account audit
TriggerEvery campaign mapped to its audience, objective and purchase signals.
02Consolidation
Logic80-plus campaigns folded into four funnels.
03Prospecting and lookalikes
AudienceNew people who resemble existing buyers.
04Retargeting and high intent
AudienceVisitors, engagers, carts and past buyers, kept separate.
05Pixel signals
DataView, add to cart and purchase events feed each funnel.
06Budget consolidation
LogicSpend follows the funnel with the strongest purchase signals.
07Purchases
ResultMore orders on a budget held at the same level.
Our approach
How we worked.
Audit
Mapped all 80-plus campaigns and the audiences they overlapped.
Design
Defined the four funnels and which audiences belong to each.
Rebuild
Consolidated campaigns and separated audiences without pausing sales.
Measure
Compared the same ten-day window before and after the change.
Run
Continuous monitoring by funnel stage, with the structure kept simple.
The outcome
What changed.
Four funnels replaced more than 80 campaigns, and each stage can now be judged on its own.
In the same ten-day window a month earlier, the account recorded 373 results across mixed objectives on $2,240 of spend. After the restructure, the same window produced 643 purchases on $2,890.
Cost per purchase settled at $4.50 after the restructure.
Fewer learning-phase resets, because campaigns are no longer fragmented.
Where this goes next
Once the campaigns bring buyers in, the next loss is the visitor who cannot find the right product. See the shopping assistant we designed for an online retailer, which recommends only what is in stock.
Retail & E-commerceAI shopping assistant that recommends from a live catalogue
Questions
What people ask.
Why do too many campaigns hurt results?
Each campaign needs enough conversions to leave Meta's learning phase. Spread a budget across dozens of campaigns and none of them gets there. Overlapping audiences also mean the account bids against itself for the same person.
Did sales drop while the account was rebuilt?
We compared the same ten-day window before and after the change, so the effect of the restructure could be seen rather than guessed. Both windows, with their spend, are shown on this page.
Is four funnels the right number for every store?
No. Four fitted this account: prospecting, lookalikes, retargeting and high intent. A smaller store might need two. The principle is the same: one campaign per job, with enough budget to learn.
