Campaign Insights

Campaign Insights is where you see how your campaigns are performing — and act on them. Every connected one shows up here; the ones an optimizer agent manages get the full dimension-by-dimension breakdown and the in-place actions, and the rest show their core metrics.

Insights start building as soon as you connect an ad account and enable it, and they're specific to the current project.

The Campaigns list

The Campaigns screen opens with summary tiles — total campaigns, how many are Zimmer-created, how many are live, total ad spend, how many are at risk, and pending proposals.

Find what you need:

• Search by campaign, agent, or platform.
• Sort by recently updated, last observed, pending proposals, spend, budget, status, platform, or name — ascending or descending — and page through the results.
• Toggle pending proposals only to see just what needs your attention.

Filter by:

Each row shows the campaign (name, objective, flight dates), platform, status, health, budget, spend, the KPI and its value, the goal it's attached to, and the agent managing it (or "Unmanaged"). It also shows when the campaign was last observed, with a "N pending" link straight to its proposals. From a row you can jump to the campaign, open it in the ad platform, or open its agent.

Health

Health is an at-a-glance read of how a campaign is doing. In the list it's a color-coded label — Critical · At risk · Monitoring · On track · No data. On a campaign's own page you'll also see a health score out of 100 (Good · Needs attention · Unhealthy), whether it's improving or slipping, and a line like "12 dimensions analyzed — 8 healthy, 3 need attention, 1 no data."

A campaign's detail page

Open a campaign for the full picture. The header shows its name and status, an Open in platform link, and a comparison control that switches what performance is measured against — your target, your peers, or your brand history. Below that, a summary band: spend, KPI, health, last observed, plus budget, flight dates, objective, and the managing agent. Then these tabs:

Every managed campaign gets this analysis on every platform. The in-place actions — the ⋯ menus that exclude a slice, adjust a budget, or pause an ad without leaving Zimmer — are rolling out on Meta first; on other platforms you'll review the same recommendations here and apply them in the platform's own UI.

Overview

• A KPI trend chart, switchable between KPI / CPM / CTR / CPC over 7, 30, or 90 days (optionally compared to the previous period). It marks the point the optimizer was connected — so you can see performance from before it took over — and the changes it has made since.
• Four cards: spend (with a daily-budget line when the campaign holds the budget), CPM (with impressions), conversions, and vs Target (how far off your KPI target you are — or the KPI value itself if no target is set).
• The budget strategy in use (for example, ABO — ad-set budget optimization) and a recent-activity list of what the agent has done.

For agent-managed campaigns, Overview also adds three panels:

• Health gauge — a score out of 100 (Good, Needs Attention, or Unhealthy) with its trend.
• Dimension Health grid — each dimension's status and winner/loser counts; click one to open it in the Dimensions tab.
• Top Signals — the top winners and losers, each linking through to Dimensions.

Dimensions

Break the campaign down by slice. Which slices appear depends on what the analyst found for the campaign and platform; common ones are ad set, audience, age, gender, device, placement, country / region, hour of week, and ad format. Most dimensions score each value across all the campaign's ad sets (the "cross-ad-set patterns" view) — for example, which age bracket wins across ad sets.

Each value's row shows:

• Pattern — Consistent Winner, Mostly Winning, Mixed, Mostly Losing, Consistent Loser, or Insufficient.
• Consistency (W/L) — a win / neutral / lose bar with counts.
• Avg Score — an index versus your benchmark (for example, +34% or 5×), green above par, red below.
• vs Target / Peer / Brand — the % delta against whichever baseline your comparison control is set to.
• Confidence — a 1–5 star rating, alongside the value's spend, KPI, and a trend sparkline.

Working with the table:

• Filter — a sidebar narrows the list by health (All / Healthy / Unhealthy / No Data).
• Sort and add columns — above the table, sort by Pattern, Score, Spend, or Consistency, and add CPM / CTR / CPC columns from the Columns picker.
• Expand a row — for a per-ad-set breakdown: each ad set's classification, score, KPI, spend, and whether a proposal was raised.
• Act in place — where in-place actions are available, a row's ⋯ menu acts on that slice, and a strong winner in a targetable dimension (country, age, gender, placement, device, or OS) gets a Create Focused Ad Set button that spins up a new ad set aimed at just that segment.

Creatives

Every creative is ranked Winning, Losing, or Neutral. The Creative Performance Map plots each one by KPI (vertical) against spend (horizontal), with bubble size showing impressions and color showing the ranking (green winning, red losing, grey neutral).

Below it, a table lists every creative — thumbnail, name and format, status, performance vs target (with its score index), spend, CTR, frequency, CPA, and a trend sparkline. (On Meta, the frequency column shows the average number of times one person saw the ad, and turns amber at 3 or more — a sign of creative fatigue.) You can sort by score, spend, or recency, filter by Winning / Losing, or search.

From each creative you can:

• Compare — pick any two creatives to see them side by side (spend, CTR, CPA, conversions), with the biggest difference called out.
• Replace — when a creative is losing or fatigued, Upload Replacement opens a drawer to swap in a new image, video, or tag (with an option to pause the original). Available on Meta, Google, TikTok, and DV360, each with its own supported formats.
• ⋯ menu — act on the ad directly, such as pausing it.

A fatigue alert flags creatives whose performance has slipped over recent runs, each with a one-click Upload Replacement.

Adsets

The ad-set view opens with four summary cards:

• Severity distribution — how the ad sets split from healthy to extreme.
• Proposal funnel — proposals → approved → executed → improved, over the last 14 days.
• Spend allocation — a treemap of where the budget is going.
• Activity strip — the last 14 days by action type (pause · budget/bid · exclude · advisory).

Below them is a table of every ad set, filterable by status. Each row shows its delivery, goal, budget, budget mode (CBO / ABO / ASC), status, severity, spend, KPI, vs target / peer / brand, confidence, impact, and trend.

From an ad set's ⋯ menu you can increase or reduce its budget (a percentage slider, capped by the agent's limits) or pause it. Click any ad set for its own detail.

Proposals

The pending proposals for this campaign, in one place. A note at the top reminds you these are scoped to this campaign — the full approval workflow (approving a whole group or run at once) lives on the agent's proposals page in Runs.

Quick filters across the top narrow the list, each with a live count:

• Tier — Tier 1 (within-ad-set, low-risk targeting tweaks), Tier 2 (entity-level budget and bid changes), Tier 3 (campaign-wide, high-impact), and Advisory.
• Action — Exclude, Boost, Bid, Budget, or Monitor.
• Confidence — High, Medium, or Low.

A summary line also calls out how many proposals sit at or above the severe threshold. Proposals are grouped by tier and then by ad set. Each one shows its action (for example, "West Java → Exclude West Java"), the KPI it targets, a confidence rating, its status (for example, Awaiting approval), a plain-language reason, and any cautions — a weak or conflicting signal, a KPI mismatch, or a change that must be applied by hand in the platform's own UI.

Act on a proposal right here: Approve or Reject it, or — when there's no automatic lever — Mark as handled once you've made the change yourself (this records who reviewed it; it doesn't push to the platform).

History

How the campaign has changed over time:

• Dimension Health timeline — healthy / unhealthy / error / no-data counts with a coverage line (it appears after a few analyst runs).
• Outcome distribution — how past changes turned out: improved, neutral, regressed, or still measuring.
• Estimated cumulative impact — spend saved, with a confidence range.
• Top actions by impact — the changes that moved the needle most.
• Action history — the full log of executed changes.

Taking action from a campaign

You can act on a campaign right here, without leaving for Runs:

• Review the agent's proposals in the Proposals tab — Approve, Reject, or Mark as handled.
• Make a change yourself with the ⋯ menus (rolling out on Meta first):
  • On an ad set — increase or reduce its budget, or pause it (with a confirmation showing the projected new budget).
  • On a creative — replace a tired one (upload a new image, video, or tag, optionally pausing the original) or pause an ad.
  • On a dimension value — exclude an underperforming slice, or spin up a focused ad set around a winning one.

To review and approve across all your agents at once, use Runs.

Metrics you'll see

Your chosen KPI is shown against your target — and, where available, against your peers and your brand history.

How it connects

• Optimizer agents act on what Campaign Insights surfaces — review and approve their changes in Runs.
• Set the targets your campaigns are measured against in Goals.
• For the cross-platform, workspace-wide rollup (built from brand training), see Brand Insights.