# Incrementality Estimator MCP > Incrementality Estimator calculates true marketing ROI by proving campaign lift above what would happen naturally. It uses control group data to establish a baseline conversion rate first, then determines the absolute and percentage gain from your advertising spend. This gives you an accurate return on ad spend (ROAS) figure that isn't inflated by organic activity. ## Overview - **Category:** analytics - **Price:** Free - **Tags:** incrementality, roas, lift, conversion-rate, control-group ## Description Standard Return on Ad Spend (ROAS) is flawed; it treats every single conversion as a campaign win. But much of that traffic would have converted anyway, regardless of your ads running. This MCP solves for that baseline noise. It establishes statistical proof of lift by comparing performance against an untreated control group. You pass in the raw data from both groups, and the system isolates the absolute activity strictly attributable to your campaign treatment. The result is a genuine marginal return figure—a true ROAS number you can use to make budget decisions. Vinkius hosts this MCP so your agent client connects once and gives you access to this critical analysis right alongside other analytics tools. ## Tools ### calculate_baseline_metrics Required before any lift calculation can begin. Calculate baseline conversion rates for control and exposed groups ### calculate_incrementality_metrics Determines how much better the exposed campaign performed versus what would have happened without it. Calculate incrementality metrics including absolute lift, percentage lift, and incremental CPA ### calculate_roas_comparison Calculate true incremental ROAS and compare it to reported channel ROAS ## Prompt Examples **Prompt:** ``` We ran a campaign for $10,000. The exposed group had 500 users (20 conversions), and the control group had 400 users (8 conversions). What is our true lift? ``` **Response:** ``` First, run `calculate_baseline_metrics` with exposed group size=500, exposed conversions=20, control group size=400, and control conversions=8. Next, use `calculate_incrementality_metrics` to find the absolute lift. Finally, if average conversion value is $50, run `calculate_roas_comparison` to get the true ROAS. ``` **Prompt:** ``` Given a baseline CR of 3% and an exposed group size of 10,000 with a total cost of $20,000. Calculate the estimated incremental conversions and resulting CPA. ``` **Response:** ``` I will use `calculate_incrementality_metrics` first to determine the absolute lift volume based on the 3% baseline. Then, I can pass that incremental volume and the $20,000 cost into `calculate_roas_comparison` to see how profitable our true return is. ``` **Prompt:** ``` Our total campaign spend was $5,000. We identified 120 truly incremental conversions with an average value of $75 each. What is the true ROAS? ``` **Response:** ``` I will use `calculate_roas_comparison` directly. I need to input incremental conversions=120, total campaign cost=$5,000, and average conversion value=$75. This gives us the most accurate profitability metric. ``` ## Capabilities ### Establish baseline metrics It calculates the natural conversion rates for both the exposed and control groups, setting a standard for comparison. ### Determine campaign lift volume The system measures how much better your ad-exposed group performed compared to the baseline—providing absolute and percentage uplift figures. ### Calculate true incremental ROAS It converts the calculated lift into a genuine Return on Ad Spend, providing an actionable profitability metric. ## Use Cases ### Budget Justification After a Campaign The Performance Manager needs to justify a $50k campaign spend. They run `calculate_baseline_metrics` first, then use `calculate_incrementality_metrics` on the results. The final ROAS comparison proves that the campaign delivered 120 truly incremental conversions, securing the next quarter's budget. ### Evaluating a New Channel A Growth Lead wants to know if a new social channel is genuinely effective or just capturing existing traffic. They feed the data into `calculate_baseline_metrics` and then use `calculate_incrementality_metrics` to see if there's any lift above historical performance. ### Comparing Campaign ROI Across Channels You need one number to compare Google Ads vs. Facebook Ads accurately. You run the full sequence: baseline, incrementality, and then `calculate_roas_comparison` for both channels. This ensures you're comparing true lift, not just gross revenue. ### Troubleshooting Low ROAS Reports The marketing team reports low ROI, but the spend was high. Instead of panicking, they use `calculate_roas_comparison` to check if their standard reporting is mixing up incremental value with baseline activity. ## Benefits - Stop basing budgets on inflated data. By running `calculate_baseline_metrics`, you establish a clear natural conversion rate for comparison. - You isolate the true gain by using `calculate_incrementality_metrics`. This reveals the absolute and percentage lift attributable only to your campaign, filtering out organic activity. - It gives you actionable financial data. The final step, running `calculate_roas_comparison`, provides a marginal ROAS figure that justifies spend. - The workflow is built for accuracy. You don't have to manually track control groups; the MCP guides the process from baseline through final comparison. - This eliminates 'vanity metrics.' Instead of just reporting ad spend vs. revenue, you report proven incremental value. ## How It Works The bottom line is you move beyond simple reporting; you quantify the exact monetary value of your intervention. 1. First, run `calculate_baseline_metrics` by supplying data from both your exposed group and the untreated control group to establish the natural conversion rate. 2. Next, feed those baselines into `calculate_incrementality_metrics`. This calculates the absolute lift volume and percentage uplift that only the campaign generated. 3. Finally, pass the incremental volume and total spend into `calculate_roas_comparison` to get a true ROAS number for decision-making. ## Frequently Asked Questions **What is the difference between reported ROAS and true incremental ROAS?** Reported ROAS treats all conversions as campaign-attributable, which usually overestimates value. True Incremental ROAS uses the `calculate_roas_comparison` tool to factor in the baseline conversion rate derived from the control group, providing a much more accurate measure of marginal return. **I only have raw cohort sizes and conversions. What should I run first?** You must start by using the `calculate_baseline_metrics` tool. This establishes the fundamental conversion rates (Control CR and Exposed CR) needed for all subsequent lift calculations. **How do I find the absolute number of extra conversions? Which tool handles this?** The `calculate_incrementality_metrics` tool is designed for this. It takes your baseline CR and exposed performance data to calculate the absolute incremental conversions, telling you exactly how many extra sales were achieved. **If I have missing or null data points when using `calculate_baseline_metrics`, how will the tool handle it?** The function requires all specified inputs (group sizes, conversions) to be numerical. If required fields are null or non-numeric, the MCP throws an error and stops processing. You must pre-validate your raw data before running this initial calculation. **What should I do if `calculate_incrementality_metrics` returns values that seem statistically unreliable?** Small sample sizes often lead to volatile lift figures. While the tool can process low volumes, we recommend analyzing groups with a minimum of 100 users each for stable and actionable incrementality metrics. **Does running `calculate_roas_comparison` require that I first run all three tools in sequence?** No, the comparison tool is robust enough to take necessary inputs directly. However, we strongly suggest using `calculate_incrementality_metrics` beforehand because it validates and optimizes your lift figures for accurate ROAS modeling. **What data format or prerequisites are needed before I can start with the initial step: `calculate_baseline_metrics`?** You need two specific sets of raw numbers: the total user count and total conversion count for both your exposed group and your control group. These counts must cover the identical time period you want to analyze. **If I run multiple comparisons in a session, are there rate limits when chaining `calculate_incrementality_metrics` with `calculate_roas_comparison`?** Vinkius manages usage rates automatically. For heavy or continuous analysis, you may encounter throttling errors. Batching your calls or scheduling the processing will help prevent hitting platform-level usage limits.