# Attribution Model Comparator MCP for AI Agents AI Agent Connect > Attribution Model Comparator lets you compare five different marketing models to see how revenue is distributed across your channels. It calculates the weight of every touchpoint so you can see which ads actually drive sales and which ones are just getting lucky. ## Overview - **Category:** marketing-analytics - **Price:** Free - **Endpoint:** https://edge.vinkius.com/vk_preview_iZ45hRxjO4K9loq9CyaUIA7eVM6WcbEsi2M1DrGa/ai-agent-connect - **Tags:** attribution, model-comparison, touchpoint-analysis, revenue-delta, marketing-mix ## Description When a customer finally buys something, figuring out which ad or email actually pushed them over the edge is a headache. You might see a sale from a social media ad, but was that just the first click, or did they need three more emails before they pulled the trigger? Traditional models often fail to capture the true value contribution across an entire customer journey. This Connector takes the guesswork out of marketing spend. It allows your AI agent to look at the entire customer journey, pull in the actual revenue from a sale, and then run it through different attribution lenses. You can see how a First Touch model weights that initial click compared to a Last Touch model that favors the final reminder. By looking at the delta between these models, you can spot which channels are getting too much credit and which ones are doing the heavy lifting behind the scenes. It's about moving past it worked to we know exactly why it worked. It covers the standard suite of models, including First Touch, Last Touch, Linear, Time Decay, and Position-Based. This gives you a clear look at how different logic impacts your reported ROI. Identifying these discrepancies is key to optimizing your budget. If one model shows a channel is a hero and another shows it is a dud, you need to know why. You can find this and thousands of other tools in the Vinkius catalog to keep your stack connected. ## Tools ### calculate_attribution_shares Gets the percentage weight for each touchpoint based on a specific model. This helps you see the relative value of every interaction. ### fetch_conversion_metrics Pulls the total revenue and currency for a unique conversion ID. It links your sales data to specific customer actions. ### calculate_attribution_delta Compares two models to show the revenue difference for a specific channel. This highlights which channels are being over-credited. ### get_customer_journey Pulls the full sequence of interactions a customer had before buying. It gives you the chronological context for every sale. ## Prompt Examples **Prompt:** ``` I need to compare how First Touch and Last Touch allocate revenue for customer CUST-123. Start by getting the journey data. ``` **Response:** ``` I'll pull the customer journey for CUST-123 to see every touchpoint. Then, I'll calculate the revenue split for both First Touch and Last Touch models. **First Touch:** $45.00 (Social Ad) **Last Touch:** $15.00 (Email Reminder) This shows that while the social ad got the first click, the email was the final driver. ``` **Prompt:** ``` Analyze the attribution shares for a high-value conversion ID CONV-900. Use all five models and target 'Google Paid Search' to see its delta. ``` **Response:** ``` I've analyzed the revenue distribution for CONV-900 across all five models. Here is how 'Google Paid Search' compares: - **First Touch:** 40% - **Last Touch:** 10% - **Linear:** 20% - **Time Decay:** 15% - **Position-Based:** 25% The delta shows that Google Paid Search is significantly over-credited by the First Touch model compared to the others. ``` **Prompt:** ``` I have the journey data and total revenue. Please calculate the attribution shares for all five models for this customer, focusing on understanding the relative contribution of Email Newsletter. ``` **Response:** ``` I've calculated the shares for all five models. Here is the contribution for the Email Newsletter: - **First Touch:** 0% - **Last Touch:** 35% - **Linear:** 15% - **Time Decay:** 20% - **Position-Based:** 25% The Email Newsletter shows its highest impact in the Last Touch and Position-Based models, suggesting it's a strong closer. ``` ## Capabilities ### Calculate revenue weights for specific touchpoints Get the exact percentage of credit for every interaction in a customer's path. ### Pull total revenue and currency for specific conversions Retrieve the final sale amount and currency tied to a specific conversion ID. ### Compare model discrepancies to find over-credited channels See the revenue difference between models to identify where spend is being misattributed. ### Retrieve chronological customer touchpoint sequences Get a full timeline of every interaction a customer had before making a purchase. ### Identify high-performing marketing channels across multiple models Analyze how different attribution logic changes the perceived value of your marketing channels. ## Use Cases ### Checking if social ads are actually driving sales A growth manager wants to know if social ads are actually driving sales or just being first touch noise by comparing models. ### Comparing Linear vs. Time Decay models A performance marketer needs to see the revenue difference between a Linear and Time Decay model for a specific channel. ### Analyzing high-value customer journeys An analyst wants to see the exact revenue weight for an Email Newsletter for a specific high-value customer using multiple models. ### Visualizing the full path to purchase A brand manager wants to see the full journey of a customer who spent over $500 to see every ad they saw before buying. ## Benefits - Identify hidden winners by using calculate_attribution_shares to see how early touchpoints influence the sale. - Spot over-credited channels with calculate_attribution_delta to see where your budget might be wasted. - Get a full view of the path to purchase by pulling chronological sequences with get_customer_journey. - Standardize your reporting by comparing five models (Linear, Time Decay, etc.) in one go. - Save hours of manual spreadsheet work by letting your agent fetch conversion metrics via fetch_conversion_metrics. ## How It Works The bottom line is you get a clear, data-backed view of which marketing channels actually deserve the credit for your revenue. 1. Provide a conversion ID to pull the full chronological customer journey. 2. Have it fetch the total revenue and currency for that specific sale. 3. Run the data through all five models to see the credit split and the delta between them. ## Frequently Asked Questions **How does Attribution Model Comparator help with marketing ROI?** It shows how different models weight your spend. By comparing them, you can see which channels are actually driving revenue and which ones are just getting credit for being the first or last click. **Can I see the full customer journey with Attribution Model Comparator?** Yes, it pulls the full chronological sequence of every interaction a customer had before making a purchase, giving you the context needed for accurate analysis. **What marketing models does Attribution Model Comparator support?** It supports the five standard models: First Touch, Last Touch, Linear, Time Decay, and Position-Based. **Can Attribution Model Comparator find over-credited channels?** Yes, it calculates the delta between different models. This helps you spot if a channel is being over-credited by one model while being undervalued by another. **How does Attribution Model Comparator handle different currencies?** It reconciles metrics by conversion ID and pulls the correct currency for the specific conversion event, ensuring your revenue totals are accurate. **What is the necessary input data to start an attribution comparison?** You must first use the `get_customer_journey` tool, providing a customer ID and conversion date. This establishes the ordered sequence of touchpoints needed for all subsequent calculations. **How do I get the financial value used as a baseline for attribution?** Use `fetch_conversion_metrics` with the specific conversion ID. This tool retrieves the total revenue amount, which serves as the single, absolute figure that all five attribution models will proportionally divide. **After calculating shares for multiple models, how do I find out which model is most biased?** The `calculate_attribution_delta` tool takes the full results from all share calculations (via `calculate_attribution_shares`) and compares them. It quantifies the revenue gain or loss for a specific channel across models, highlighting discrepancies.