Campaign Performance Analysis: A Practical Guide
Campaign Performance Analysis: A Practical Guide

Campaign performance analysis means working out whether a campaign did what it was meant to do, why, and what to change next. Do it in a fixed order: set the goal, choose the KPIs, benchmark, compare spend with outcomes by campaign, test the attribution, then decide. The slow part is rarely the thinking. It is pulling Google Ads, Google Analytics and Search Console numbers into one place, which is why we built Caybl to let you ask Claude or ChatGPT about those accounts in plain English and get answers from live data.
TL;DR
- Analysis is not reporting. A report shows what happened, analysis explains why and says what to do.
- Pick KPIs from the campaign goal, not from whatever the dashboard shows first.
- Judge every number against a benchmark: your own history, the previous period and, where you can find it, your sector.
- Compare campaigns on spend, conversions, cost per acquisition and return on ad spend side by side, then check attribution before you move budget.
- An AI assistant connected to your live accounts shortens the data gathering, but you still verify anomalies before acting.
Table of Contents
- What campaign performance analysis is, and what it is not
- Step 1: Start with the business goal, then choose KPIs
- Step 2: Set benchmarks before you judge anything
- Step 3: Gather the data from every source
- Step 4: Compare campaigns side by side
- Step 5: Check attribution before you move budget
- Step 6: Turn findings into decisions
- Mistakes that quietly distort the read
- Where AI helps, and what to verify
- Running this analysis with Caybl
- Our view: speed is what makes analysis a habit
- FAQ
- Sources
What campaign performance analysis is, and what it is not
A report organises data into an accurate snapshot: spend was this, clicks were that. Analysis goes further. It looks for patterns, adds context and informs a decision. Epsilon’s explainer makes the same distinction, describing reporting as the snapshot and analytics as the interpretation.
In practice the difference shows up in the questions you ask. A report answers “how much did we spend last month?” Analysis answers “why did cost per lead rise by a third in the second half of the month, and which campaign caused it?”
This guide is for people who run or review paid campaigns: marketers, business owners running Google Ads, in-house teams and agencies reporting to clients. It assumes you have the data already and want a repeatable way to read it.
Step 1: Start with the business goal, then choose KPIs
The IAB’s November 2024 cross-channel measurement playbook begins measurement with business objectives, stakeholder alignment and KPIs tied to those objectives. That order matters because a KPI chosen first tends to become the goal by accident.
A simple mapping works for most paid campaigns:
| Campaign goal | Primary KPIs | Supporting metrics |
|---|---|---|
| Awareness | Reach, impressions | Brand mentions, branded search growth |
| Lead generation | Leads, conversion rate, cost per lead | Click-through rate, lead quality |
| Sales or revenue | Revenue, conversion rate, return on ad spend (ROAS) | Average order value, customer lifetime value |
The goal-to-metric pairings follow the approach in Smartsheet’s campaign measurement guide. ROAS is revenue divided by ad spend, so £3.50 back for every £1 spent is a ROAS of 3.5.
If your campaigns also carry brand objectives, brand metrics and how to use them covers what to measure alongside the performance numbers.
Write the goal and its two or three KPIs at the top of the analysis before you open any account. If a number does not serve the goal, leave it out of the main read.
Step 2: Set benchmarks before you judge anything
A cost per acquisition of £50 is neither good nor bad on its own. The IAB playbook recommends setting internal and industry benchmarks, both for overall cross-channel performance and for individual channels or campaigns. Use three reference points:
- Your own baseline. The average of the last three to six months for that campaign type.
- The previous period. Month on month, plus year on year where seasonality matters. The Upward Engine post-campaign guide recommends both, so you can tell a longer-term trend from a short-term wobble.
- A target. The CPA or ROAS the business needs to be profitable, which comes from margin, not from the ad platform.
The third is the one most often skipped. A campaign can beat its benchmark and still lose money if the benchmark was never tied to margin.
Step 3: Gather the data from every source
For a typical paid campaign you need three views of the same activity:
- Google Ads for spend, impressions, clicks, conversions and cost per conversion by campaign, ad group and keyword.
- Google Analytics for what visitors did after the click: engagement, conversions, revenue by channel.
- Search Console for organic clicks, impressions and ranking changes, which explain traffic shifts that paid data cannot.
The IAB playbook suggests auditing your existing tools, data sources, integration needs and team capability before building measurement. In plain terms: list where each number lives, who owns it, and whether two tools define “conversion” differently. Mismatched definitions cause more arguments in review meetings than genuine performance problems.
This is where most of the time goes. Exporting from each tool, lining up date ranges and pasting into a spreadsheet can take an afternoon, and the sheet is stale by the following Monday. For more on what to connect and why, see our guide to marketing platform integrations.
Step 4: Compare campaigns side by side

Once the data is in one place, put campaigns in a single table with the same columns: spend, conversions, CPA, revenue and ROAS. The Upward Engine guide recommends exactly this side-by-side comparison of spend, conversions, CPA and ROAS.
Here is an illustration with made-up figures:
| Previous month | This month | Change | |
|---|---|---|---|
| Spend | £3,500 | £4,000 | +14% |
| Conversions | 85 | 80 | -6% |
| CPA | £41.18 | £50.00 | +21% |
| Revenue | £13,600 | £14,000 | +3% |
| ROAS | 3.89 | 3.50 | -10% |
The headline looks fine: revenue grew. The analysis says otherwise. Spend rose faster than revenue, so each extra pound bought less. The next question is where the extra £500 went. Break the table down by campaign. You may find one broad-match campaign took most of the increase while a branded campaign held steady.
Two habits keep this honest:
- Compare like with like. A prospecting campaign and a brand-search campaign have different jobs. Judging them on the same CPA hides the real picture.
- Look at the tail. Sort by spend and check the bottom third. Small campaigns and keywords with spend and no conversions are where wasted budget usually sits.
Our guide to paid search reporting lists what to track for search campaigns specifically, and the ecommerce reporting guide does the same for shops with a weekly routine.
Step 5: Check attribution before you move budget
Attribution decides which touchpoint gets credit for a conversion. The IAB UK measurement toolkit describes it as assigning KPIs and conversion measurement across the customer journey, from brand-equity measures through to sales. It also says attribution can give actionable short-term campaign insight and complements higher-level marketing mix modelling.
The IAB playbook advises choosing models that fit your goals and the complexity of the customer journey, possibly combining single-touch, multi-touch and data-driven approaches, and testing and refining them over time.
What this means for you: before you cut a campaign for a poor last-click ROAS, check whether it starts journeys that other channels finish. Compare the same campaign under two attribution models. If it looks very different, the decision needs more evidence. Our cross-channel attribution guide walks through models, setup and validation.
Step 6: Turn findings into decisions
Analysis ends with an action list, not a chart. For each finding, write one line in this shape: what changed, the likely cause, the action, the owner and the date you will check again.
For the illustration above, that could read: CPA up 21% month on month, driven by a broad-match campaign; tighten match types and add negative keywords; owner is the account manager; review in two weeks.
The Upward Engine guide ends with an action plan and recurring reviews, and that second part matters. One-off analysis fixes one month. A fixed weekly or monthly slot catches drift early. If you report to clients or a board, keep the format the same each time so changes stand out. Our guides to marketing dashboards and data visualisation show how to present the result so decision-makers read it.
Mistakes that quietly distort the read
- Reading platform conversions as final truth. Ad platforms count conversions their own way. Reconcile with Analytics or your sales data.
- Judging on too little data. A campaign with a handful of conversions swings wildly. Wait for enough volume or widen the date range.
- Ignoring seasonality. A December dip against November may be the calendar, not the campaign. Use year-on-year views.
- Mixing goals. Comparing awareness and sales campaigns on the same KPI punishes the wrong one.
- Forgetting organic. A traffic drop might come from lost rankings, not ad changes. Search Console settles that quickly.
Where AI helps, and what to verify
The IAB playbook recommends predictive analytics and machine learning to identify patterns and forecast trends, alongside monitoring data for trends, gaps and anomalies, investigating issues and regular quality assurance. That last part is the human job. An assistant can surface an anomaly in seconds, but you decide whether it is a tracking fault, a seasonal effect or a real problem.
A sensible routine when you use an AI assistant on your accounts:
- Ask for the comparison and the numbers behind it, not just a conclusion.
- Spot-check one or two figures against the platform’s own interface.
- Preview any proposed change before it goes live.
For a wider view of using AI across the marketing workflow, see our practical guide to using AI for marketing.
Running this analysis with Caybl
Caybl connects Google Ads, Google Analytics and Search Console to Claude and ChatGPT, so the data-gathering steps above become questions. Ask which campaigns are wasting spend, why traffic dropped last week or which pages lost rankings, and the answer comes from your live data, with no exports, no spreadsheets and no dashboard to learn.
Setup takes about two minutes with a Google sign-in. Access is scoped to what you approve, changes are previewed before they apply, and you can disconnect in one click.
The Free plan costs $0 and Google Analytics is free to connect. Pro is $29 per month and adds Ads, Search Console and more, and Team starts from $20 per month per seat for agencies and marketing teams. See the Caybl pricing page for the full list, or read how to connect Google Ads to Claude.

Our view: speed is what makes analysis a habit
Most teams know what good campaign analysis looks like. They set goals, compare periods, check attribution. What stops them is the gathering: logging in to three tools, exporting, matching date ranges, rebuilding the same sheet every week. So the analysis happens monthly at best, and by then the money is spent.
We built Caybl on a simple idea: you should be able to ask your own accounts a question in plain English and get an answer from live data. Which campaigns are wasting spend? Why did traffic drop last week? Which pages lost rankings? When the answer takes a minute rather than an afternoon, you ask more often, and the weekly review stops being a chore.
We also care about control. Access is scoped to what you approve, changes are previewed before they apply, and you can disconnect in one click. Analysis should make you quicker, not make you nervous about what happens to your accounts.
Gearoid
FAQ
What are the main steps in campaign performance analysis?
Set the business goal, choose KPIs that match it, set benchmarks, gather data from every source, compare campaigns side by side, check attribution and record actions with owners. The IAB playbook follows the same logic: goals and KPIs first, then benchmarks, data, attribution and ongoing monitoring.
Which metrics should I use to analyse a campaign?
Let the goal decide. Awareness campaigns suit reach and impressions, lead generation suits leads, conversion rate and cost per lead, and sales campaigns suit revenue, conversion rate and ROAS. Keep two or three primary KPIs and treat the rest as supporting context.
How often should I review campaign performance?
Review active paid campaigns weekly for spend and conversion drift, and run a fuller comparison monthly, with year-on-year views where seasonality matters. Recurring reviews catch problems earlier than a single end-of-campaign analysis.
Can I analyse Google Ads and Analytics data with Claude or ChatGPT?
Yes. Caybl connects Google Ads, Google Analytics and Search Console to both assistants through a Google sign-in, so you can ask questions about your own accounts in plain English. Google Analytics is free to connect, and our guide shows how to connect Google Analytics to ChatGPT.
Sources
- IAB: Implementing Cross-Channel Measurement for Marketers Playbook, November 2024
- IAB UK: Measuring Digital Advertising in a Multi-Media Context
- Epsilon: What is the difference between reporting and analytics?
- Smartsheet: How to Measure Marketing Campaign Effectiveness
- Upward Engine: Post-Campaign Analysis Step-By-Step Guide
- Brand Metrics: What to Measure and How to Use Them