> For the complete documentation index, see [llms.txt](https://docs.roivenue.com/roivenue-resources/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.roivenue.com/roivenue-resources/roivenue-training/roivenue-app/attribution/model-comparison.md). # Model Comparison Compare how attribution models credit your channels and campaigns. Put a base model next to up to two other models and see where they agree and where they differ. Model Comparison shows how your results change when you switch attribution models. Every model splits the same conversions and revenue differently. Last Touch gives all credit to the last click. AI Data Driven spreads it across the whole customer journey. This page puts the results side by side, so you see where the models disagree and which channels your current reporting over- or under-credits. Use Model Comparison when you want to: * compare the numbers you know, usually Last Touch or Platform Data, with AI Data Driven, * find channels and campaigns that start or support customer journeys but get little credit from the last click, * check how much revenue sits in rows like **Direct** or **Organic** under each model, * explain to colleagues why Roivenue numbers differ from GA4 and from the numbers in the ad platforms. The biggest differences usually point to the biggest optimization potential. The page has four parts: 1. **Header:** Share, Save, Attribution Settings, currency selector, and date range. 2. **Table setup:** dimensions, measures, and attribution models for the comparison, plus table actions. 3. **Comparison table:** results of the base model, the compared models, and the differences between them. 4. **Filter Panel:** filters for the whole table. When you close it, open it again with the filter icon in the header. To open the page, go to **Attribution → Model Comparison** in the left navigation. The date range and the **Filter Panel** apply to the whole table. See [Date Picker](https://docs.roivenue.com/roivenue-resources/roivenue-training/roivenue-app/date-picker) and [Filter Panel](https://docs.roivenue.com/roivenue-resources/roivenue-training/roivenue-app/filter-panel).
*** ### How to set up the comparison table The bar above the table has four settings. | Setting | What it does | Limit | | ------------------------- | --------------------------------------------------------------------------------------------------------- | ------- | | **Dimension** | Defines the rows. The order sets the drill-down, for example **Platform Name**, then **Campaign (auto)**. | Up to 3 | | **Measure** | The metric you compare, for example **Net Revenue** or **Conversions**. | Up to 2 | | **Base Attribution** | The reference model. All differences are calculated against it. | 1 | | **Compared Attributions** | The models you compare with the base model. | Up to 2 | Click a setting to open its panel: * Add an item with **Add Item**. The list has a search field. * Change the order by dragging the handle on the left. * Remove an item with the bin icon. When you reach the limit, **Add Item** disappears.
The model lists contain the attribution models of your account, for example, **Last Touch**, **First Touch**, **Linear**, and **AI Data Driven**. They also include **Platform Data**, which is what the advertising platforms report themselves, and the default UMM or MMM model if your account uses them.
To see or change the attribution models of your account, click **Attribution Settings** at the top of the page. It opens **Account Configuration → Attribution Models**. {% hint style="info" %} **Tip:** Set **AI Data Driven** as the Base Attribution. It is the Roivenue model, so you compare all other models against it, for example **Last Touch** and **Platform Data**. As dimensions, use **Platform Name** and then **Campaign (auto)** if you want to see the campaign level and Campaign (auto) works in your account. See [Campaign (auto) Dimension](https://docs.roivenue.com/roivenue-resources/roivenue-training/core-concepts/campaign-auto-dimension). As the measure, use **Net Revenue** if you work with conversion values. Use **Conversions** if you run a lead-based business without conversion values. {% endhint %} *** ### How to read the comparison columns For each measure, the table shows these columns: 1. **Base model**, for example AI data driven. 2. **Compared model**, for example Last touch. 3. **+/-**: the absolute difference between the compared model and the base model. 4. **%**: the relative difference. The difference headers name both models by their short names, for example **+/- (LT vs. AI DD)**. Short names come from **Attribution Settings**. Platform Data uses **PD**.
Green cells mean the compared model gives the row more credit than the base model. Red cells mean less credit. The stronger the color, the bigger the difference. Example from the screenshot: AI Data Driven credits Google Ads with EUR 10.1m of Net Revenue, while Last Touch credits it with EUR 5.4m. The difference is EUR −4.7m, or −46.6%. Last Touch misses the role Google Ads plays earlier in the customer journeys. Direct shows the opposite: Last Touch gives it EUR 6.6m more than AI Data Driven, because the last visit before a purchase is often direct. The **Cost** column is the same for all models, because attribution splits only conversions and revenue, not spend. The **Total** row at the bottom sums the cost and each model's measure. {% hint style="info" %} **Tip:** Columns are narrow when you compare two models. Close the Filter Panel to get more space, or hover over a column header to read its full name. {% endhint %} *** ### Why Last Touch and Platform Data differ from GA4 and the ad platforms Model Comparison often shows numbers that don't match what you see in GA4 or in the ad platforms. This is expected. Each tool counts the conversions differently. | Comparison | Why the numbers differ | | ------------------------------------------------ | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | **Roivenue Last Touch vs. GA4** | GA4 uses last non-direct click. Direct traffic gets credit only when it is the only touchpoint. Roivenue Last Touch includes direct visits, so **Direct** gets more credit. For the path Display → Social → Paid Search → Direct, GA4 gives 100% to Paid Search and Roivenue Last Touch gives 100% to Direct. A difference of around 20% is normal. | | **Platform Data vs. Roivenue models** | Platform Data are the conversions each platform reports about itself. A platform sees only its own campaigns and credits itself with the whole conversion, often the same one another platform claims too. Roivenue models count each conversion once and split it between the touchpoints, so the Platform Data total is usually higher. | | **Platform Data vs. the platform's own reports** | The numbers should match when you compare the same conversion type, the same attribution window, and the same reporting date. The attribution window is set in the connector or fixed by the platform's API, so it can differ from the settings in the platform's reports. | Other differences from GA4 come from the data itself. With Roivenue Measurement, Roivenue also tracks visitors that ad blockers or consent settings hide from GA4. With a Transaction Feed, it counts orders from your order system, including orders GA4 missed. Platform Data don't enter any attribution model. Roivenue uses them only to create synthetic impressions. See [Last Touch vs. AI Data-Driven Attribution](https://docs.roivenue.com/roivenue-resources/roivenue-training/core-concepts/last-touch-vs.-ai-data-driven-attribution) and [Attribution Window](https://docs.roivenue.com/roivenue-resources/roivenue-training/core-concepts/attribution-window). *** ### What the icons in the rows mean
1. **Eye icon (Impressions)** and **cursor icon (Clicks)** show which touchpoints are part of the customer journeys for this row. * Impressions can be pixel-tracked impressions from Roivenue Measurement or synthetic impressions that Roivenue creates from the conversions the platforms report. Both let models like AI Data Driven credit view-based activity. See [Click-based vs. Impression-based Touchpoints](https://docs.roivenue.com/roivenue-resources/roivenue-training/core-concepts/click-based-vs.-impression-based-touchpoints). * A row with only the cursor icon has no impressions in its journeys. 2. **Cost icon:** the row has no cost data. Hover over it to see what to do. * **Add an expense** opens **Manual Costs**, where you upload costs for sources without a connector. * **Create a connector** opens **Connectors**, because Roivenue has a connector for this platform. The number in brackets after the row name shows how many items the row groups at the next dimension level. For example, **Google Ads (76)** contains 76 campaigns. *** ### How to drill down and work with the table When you use more than one dimension, click the arrow next to a row to see the next level, for example the campaigns of one platform.
Every column header has a filter icon and a ⋮ menu (1). Hover over the header to see them. The menu (2) has these options:
* **Sort Ascending** and **Sort Descending** sort the rows by the column. You can also click the header itself. * **Pin Column** keeps the column visible when you scroll the table sideways. * **Autosize This Column** fits the column width to its content. * **Choose Columns** lets you hide or show columns. You can also: * filter a column with the filter icon in its header, for example, to show only rows with a value greater than or equal to a number, * drag the edge of a column header to make the column wider, * change the number of rows with **Rows per page** and move between pages at the bottom of the table. The three icons on the right of the setup bar are table actions. Hover over an icon to see its name.
IconActionWhat it does
1Expand All / Collapse AllOpens or closes all rows at once. When at least one row is open, the icon switches to Collapse All.
2ExportDownloads the table, so you can work with it in a spreadsheet.
3Add to DashboardAdds the table to a dashboard with its current settings. Roivenue names the widget after the models, for example "Last Touch vs. AI Data Driven". See Dashboard.
*** ### How to save and share the view Click **Save** in the header at the top of the page, next to **Share**.
Then choose how to save the view.
* **Save As Bookmark** keeps the page exactly as you see it, including filters and date range. You open it from **Bookmarks** in the left navigation. See [Bookmarks](https://docs.roivenue.com/roivenue-resources/roivenue-training/roivenue-app/bookmarks). * **Save As Default** makes the current table setup the default view of the page. Filters and date range are not saved. * **Share** copies a link to the current view to your clipboard. *** #### FAQ
1. Why does Platform Data in Roivenue differ from the numbers in the ad platform? The most common causes are: * **Different attribution window.** Roivenue uses the window set in the connector, for example 28-day click and 1-day view for Meta. If the platform view uses 7-day click, the numbers differ. Check the connector settings. * **Conversion date vs. interaction date.** Roivenue shows platform conversions by the date of the conversion. Google Ads shows conversions by the date of the interaction by default. Add the "by conv. time" columns in Google Ads to compare the same thing. * **Duplicate connectors.** Two connectors for the same ad account double the data. Switching one off is not enough, because the data it already downloaded stays. Delete the duplicate connector and ask Roivenue support to reprocess the data. * **Later updates in the platform.** Platforms keep changing recent numbers. Roivenue reloads the last days regularly, but a small difference can remain.
2. Why does AI Data Driven give a channel much more or less credit than Last Touch? Last Touch credits only the last click. AI Data Driven credits every touchpoint in the journey, including impressions. Channels that start or support journeys, such as video or display, usually gain credit. Rows like **Direct** or **(not tracked)** usually lose credit because AI Data Driven passes it to the touchpoints that brought the customer earlier. See [Last Touch vs. AI Data-Driven Attribution](https://docs.roivenue.com/roivenue-resources/roivenue-training/core-concepts/last-touch-vs.-ai-data-driven-attribution).
3. Why does the difference between AI Data Driven and Platform Data change from month to month? It depends on how your customer journeys look. When users convert more than once, AI Data Driven spreads every conversion across the whole journey of that user, including touchpoints before earlier conversions. A platform reports only its own conversions. Months with many repeated conversions therefore show a bigger difference than months with mostly new customers.
4. Why is the eye icon missing next to a channel or campaign? The row has no impressions in its journeys. Typical reasons: * The dimension rule for the row does not include synthetic impressions. This happens with custom dimensions. Update the rule in **Data → Dimensions**. * The platform reports very few conversions. Synthetic impressions are created from platform-reported conversions, so without them Roivenue has nothing to pair. * Your web data comes from the GA4 API, which provides only the date of a conversion, not the time. Synthetic impressions need an hourly timestamp. Connecting GA4 through BigQuery usually solves it.
5. Why do I see impressions under "Other"? The traffic source of the campaign does not match any rule of the dimension. For example, Meta campaigns tagged with the UTM source "fb" fall under "Other" when the platform rule expects a different value. Add the value to the dimension rule in **Data → Dimensions** and the impressions move to the right row.
6. Why don't First Touch results include impressions? Synthetic impressions are not created for First Touch models. First Touch therefore credits the first click in the journey. To see the effect of impressions, compare AI Data Driven with Last Touch or Platform Data.
7. Can I compare more than three models or use more than three dimensions? No. The table shows one base model and up to two compared models, up to three dimensions, and up to two measures. For another combination, change the settings or save a second setup as a bookmark.
8. Why did the numbers for past months change? Roivenue reprocesses data after changes in connectors, dimension rules, or attribution settings, and after fixes on the Roivenue side. Reprocessing recalculates historical data, so past results can change. If the change surprises you, contact Roivenue support.
9. Why does the page open with a different setup than I used last time? The page opens with its default setup. To keep your dimensions, measures, and models, click **Save → Save As Default**. To keep filters and date range too, use **Save As Bookmark**.