> 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/umm/umm-models.md).

# UMM Models

Create UMM models and set for each site how MTA and MMM are mixed: mixing metric, totals, driver weights, constraints and experiments.

A UMM model defines how Roivenue mixes your MTA data with your MMM data. You can create more models, for example to compare different weights, and see each of them in UMM Analysis.

Use UMM Models when you want to:

* create your first UMM model
* change the weight between MTA and MMM
* set different rules for a specific period
* limit how much UMM can change MTA values
* add experiment results to a model

The page has 2 levels:

1. **UMM Models:** the list of your models with **Name** and **Short Name**, and **Add UMM Model**.
2. **UMM Model Settings:** **General Settings** shared by all sites, and **UMM Site Settings** with a configuration for each site.

To open the page, go to **UMM → UMM Models** in the left navigation.

***

### How to create a UMM model

1. Click **Add UMM Model**.
2. Fill in the general settings:

| Setting                                      | What it does                                                                                                 |
| -------------------------------------------- | ------------------------------------------------------------------------------------------------------------ |
| **Name UMM Model**                           | The name of the model. It must be unique.                                                                    |
| **UMM Model Short Name**                     | A short name used where space is limited, for example in Model Comparison. Max 6 characters, must be unique. |
| **Default Data Driven Model for MMM Mixing** | The MTA model that is mixed with your MMM data.                                                              |
| **Set Default Weight Between MTA:MMM**       | The default weight for drivers without their own setting.                                                    |

3. Save the model.
4. Add a configuration for each site (see below).

To change the general settings later, open the model and click **Edit** in **General Settings**.

***

### How to configure a site

Open the model. **UMM Site Settings** lists your sites with two checks:

* **Data Uploaded:** MMM data for the site exists.
* **Data Categorized:** all drivers for the site have a category.

Both must be green before you can add a configuration. Then click **Add Configuration** (or **Edit Configuration**) and set:

| Setting                     | What it does                                                                                                           |
| --------------------------- | ---------------------------------------------------------------------------------------------------------------------- |
| **Choose Mixing Metric**    | Whether UMM is calculated from **Conversions** or **Net Revenue**.                                                     |
| **Choose Mixing Algorithm** | The algorithm used to mix MTA and MMM.                                                                                 |
| **UMM Totals**              | Whether UMM totals match **MTA** or **MMM**. With **MMM**, only the metrics from the imported MMM data are calculated. |

***

### How to set the drivers

Under **Custom Driver Settings**, each driver category has its own settings:

| Category              | Settings                                                                                                                                                                                                                                                                                            |
| --------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| **Online Media**      | **Set the weight between MTA and MMM results** for each driver. The two weights always add up to 100%. **Constrain increase to MTA evidence** stops UMM from raising the driver above what MTA supports. **Constrain decrease to MTA evidence** stops UMM from lowering it below what MTA supports. |
| **Non-digital Media** | **Target % from MMM** sets the driver's target in UMM as a percentage of its value in MMM.                                                                                                                                                                                                          |
| **Other Variables**   | No settings. Their values are added to the UMM result without modelling.                                                                                                                                                                                                                            |
| **Base**              | No target. The base is adjusted to account for changes in other drivers.                                                                                                                                                                                                                            |

Turning a constraint on keeps the result closer to MTA. Turning it off gives the model more flexibility to reach the target.

***

### How to limit changes to MTA values

By default, UMM can change touchpoint values freely. Under **MTA Value Change Constraints**, click **Add A Constraint** to limit the change:

1. Set the **Rule** that selects the touchpoints.
2. Fill in the minimum and/or maximum as a percentage of the original value, for example 70% to 180%. At least one of them is required.

{% hint style="info" %}
If a constraint is reached, the algorithm stops at the limit and the target value may not be achieved.
{% endhint %}

***

### How to add experiments to the model

Under **Experiments**, click **Add An Experiment**:

1. Select an experiment. It must use the same site as the configuration. If the totals come from MMM, its metric must match the mixing metric.
2. Set the weight between the experiment and UMM.

Create experiments first on the [Experiments](https://docs.roivenue.com/roivenue-resources/roivenue-training/roivenue-app/umm/experiments) page.

***

### How to set rules for a specific period

A **Period Rule** is a configuration that applies only for a date range and overrides the default configuration during that time.

1. Make sure the site has a default configuration.
2. Click **Add Period Rule**.
3. Set **Date Range**. Fill in at least one of **From** and **To**.
4. Set the configuration and save.

***

#### FAQ

<details>

<summary>1. Why can't I add a configuration for a site?</summary>

The site has no uploaded MMM data, or its drivers are not categorized. Check **Data Uploaded** and **Data Categorized**, then fix it in [MMM Data Upload](https://docs.roivenue.com/roivenue-resources/roivenue-training/roivenue-app/umm/mmm-data-upload) or [Driver Categorization](https://docs.roivenue.com/roivenue-resources/roivenue-training/roivenue-app/umm/driver-categorization).

</details>

<details>

<summary>2. Why can't I add a Period Rule?</summary>

A Period Rule needs a default configuration for the site. Create it first with **Add Configuration**.

</details>

<details>

<summary>3. Should UMM totals match MTA or MMM?</summary>

Choose **MTA** when you want UMM totals to stay in line with the conversions or revenue Roivenue measures. Choose **MMM** when you want them in line with your MMM model. With **MMM**, only the metrics from the imported MMM data are calculated.

</details>

<details>

<summary>4. Why is the result different from the target I set?</summary>

A constraint was probably reached. When a constraint is reached, the algorithm stops at the limit and the target may not be achieved.

</details>

<details>

<summary>5. Why can't I add an experiment to the model?</summary>

No experiment exists yet, or its site or metric does not match the configuration. Create or edit the experiment on the [Experiments](https://docs.roivenue.com/roivenue-resources/roivenue-training/roivenue-app/umm/experiments) page.

</details>

<details>

<summary>6. When do I see the results?</summary>

After you save, the model recalculates. This usually takes a few hours. The results appear in [UMM Analysis](https://docs.roivenue.com/roivenue-resources/roivenue-training/roivenue-app/umm/umm-analysis) once it is done.

</details>

<details>

<summary>7. What is the short name for?</summary>

It is used where space is limited, for example in Model Comparison. It can have at most 6 characters.

</details>
