Instant Ad Account Diagnostics: Testing mureo Without Any Data Export

Introduction

Imagine being able to run a full strategy-grounded diagnosis on a realistic ad account without exporting a single spreadsheet or granting OAuth access. That's exactly what mureo v0.8.0 delivers with its new demo init command. In under 60 seconds, you can explore how an LLM agent analyzes synthetic yet lifelike ad accounts—spotting hidden anomalies that aggregates would miss—all powered by your own strategic playbook.

The Problem: Trying Before You Invest Time

Earlier versions of mureo introduced a BYOD (Bring Your Own Data) mode: drop a Google Ads or Meta XLSX into the tool and receive a diagnosis grounded in your business strategy. While the setup only takes about five minutes, many potential users hesitated. "I don't have a Sheet export ready yet—can I just see what the output looks like first?" was the most common feedback across developer forums and direct messages.

That five-minute barrier, though small, was enough to stop people from experiencing the value firsthand. The team realized that providing an instant, zero-effort demo would help users decide whether mureo fits their workflow without any upfront commitment.

Introducing mureo demo init

With v0.8.0, a new command mureo demo init --scenario <name> materializes a complete synthetic account environment in seconds. It generates:

• A realistic XLSX bundle (simulating 90 days of ad data)
• A STRATEGY.md file that defines your business objectives
• A pre-populated STATE.json with imported account state
• Configuration files (.mcp.json, README.md)

After running the command, open the directory in Claude Code and simply ask /daily-check. The agent will reason over the synthetic account as if it were real, applying the strategy defined in your STRATEGY.md to uncover insights.

How It Works

The demo uses the exact same import pipeline as the real BYOD workflow. There is no separate demo code path. The numbers the agent analyzes come from the same .mureo/byod/ database that real users rely on. This means the experience is identical—only the data source is synthetic.

Available Scenarios

Four distinct scenarios ship with v0.8.0. You can list them with mureo demo list:

1. seasonality-trap (default) – A Meta ad account where a CPA spike appears seasonal but is actually caused by a broken Pixel after a Shopify migration.
2. silent-outlier – A B2B SaaS account where headline numbers look healthy, but a single long-tail search term quietly converts at 4x the ad group average.
3. Two additional scenarios covering common ad performance puzzles (described at the end of this article).

Deep Dive: Two Realistic Scenarios

Understanding why these scenarios matter is best done by walking through them. Below we examine the two most illustrative cases in detail.

The Seasonality Trap (D2C Cosmetics)

In this scenario, you're analyzing a cosmetics brand called FlavorBox. The Meta CPA chart looks clean for the first 21 days, then on Day 22 it spikes vertically. Over the next 25 days, the manager tries three escalating actions (pausing underperformers, raising budgets, changing audiences) but nothing bends the curve.

A dashboard-only view would suggest "seasonality" or "market downturn." But the mureo agent, grounded in your STRATEGY.md (which emphasizes high LTV customer acquisition and brand awareness), digs deeper. It flags that the spike coincides with a Shopify migration on Day 20. The broken Pixel means the platform is optimizing for low-intent actions instead of purchases. The agent recommends fixing the Pixel and using offline conversions to retrain the algorithm—a diagnosis that aggregates would miss entirely.

The Silent Outlier (B2B SaaS)

Now consider a B2B SaaS account with healthy headline metrics: CPA below target, impression share stable, ROAS acceptable. Yet within the ad groups, a single long-tail search term has a cost-per-conversion that's 4x lower than the rest. The vanilla dashboard shows the aggregate as good. The mureo agent identifies this outlier and cross-references it with the strategy document (which may prioritize high-intent niche segments). It then recommends shifting budget to exploit that term, expanding match types defensively, and creating a dedicated ad group to maximize the hidden opportunity.

Both scenarios end at the same conclusion: dashboards show aggregates, but business judgment lives in the outliers. An LLM grounded in your written strategy becomes a meaningfully different reader of those outliers compared to a vanilla LLM that only sees raw numbers.

Beyond the Demo

The two additional scenarios—one focused on budget cannibalization across campaigns and another on misattributed offline conversions—extend the same principle. They are designed to show how specific strategic contexts change the interpretation of data.

The mureo demo init feature is available now via pip install mureo (v0.8.0, released May 2, 2026). It requires no Google/Meta account, no spreadsheet export, and no OAuth token. Just run the command, open in Claude Code, and ask /daily-check. You'll see the agent reason over a realistic 90-day account in under one minute.

To learn more about the BYOD approach or explore the full mureo documentation, visit the official repository (internal link placeholder).