# Voluum Documentation

#### Optimization Calculator: Empowering Conversion Rate of Offers

Voluum TRK TRK Campaign Optimization

This article presents how the Bayesian statistical approach might give you the insight to optimize your campaign funnel by changing weights of the offers in a campaign configuration. The Optimization Calculator based on a Bayesian model introduced to Voluum enables you to take a look at estimations of different entity slots compared together in one diagram.

To find out more about the concept of the Optimization calculator in Voluum, go to the Optimization Calculator: What Converts More Likely article.

For example, you might want to find out which browsers for the particular offer within one campaign funnel might bring better results in terms of conversions. This guide instructs you how to verify the statistical estimations and apply them to an existing campaign configuration. However, you should keep in mind that it only describes one of the possible optimization where the Bayesian statistic might be incorporated.

### Caution

Voluum Note: Before you start analyzing estimations based on a Bayesian approach, you should take into account the following points:

1. In the Voluum platform at least one active campaign should be running where there are samples of data with visits and conversions; the more registered tracking events, the better; you need to keep in mind that this is a statistical approach where more data will generate more adequate estimations.

2. In the campaign funnel there is at least one offer defined.

3. You are familiar with the Voluum reporting data for the campaign funnel.

4. The Bayesian statistical estimations can shed some light how the offers might convert in the future, but the estimated numbers cannot be taken for granted.

5. Voluum reports cannot differentiate a display of data for the same offer added to different paths in one campaign. Thus, some results such as offer conversions cannot be split into single samples of data for only one path.

6. It is highly recommended to take a look at a timestamp of visits and corresponding conversions registered for your campaign funnel in Voluum. This allows you to find an average time range between recorded tracking events. It is worth comparing estimations for this time range as well as verify how they might change in a long-term period what enables you to visualize how the data varies on a timely basis.

##### I. Take a Look at the Campaign Data

Once you have set the campaign funnel and noticed that the tracking events were recorded, you can take a closer look at your data collected in Voluum:

2. Go to the Campaigns view.

3. In the Campaigns view double-click a campaign where you want to analyze your current results and check estimations of future results. The report for the selected campaign will open in a separate tab.

4. Take a look at the date/time range for which the report has been generated and verify whether it is equivalent to the time range when the visits and their corresponding conversions are registered. For example, if the visits and most of the corresponding conversions are registered within one day, you should select the Yesterday option as the time range for the Campaign report. If the conversions are registered in a very specific time range, you can always select the Custom day range or Custom time range option.

5. Starting from this point you can try to compare the Bayesian estimations for different tracking options available in Voluum. For example, you can estimate on which devices your ad might convert best based on collected statistics. Or check the probability of conversion if the offer will be displayed in different countries.

The next steps will show you how to analyze statistical estimations comparing data of the offer(s) displayed in different browsers.

1. In the Campaign report view double-click the Browsers option in the menu bar. The Campaign report reloads and you should see a list of browsers for which the data has been collected.

2. Select the checkbox on top of the the table to choose all browser entities or choose at least one entity by selecting a checkbox next to it. The Optimization calculator button will show up in the menu above the report table.

3. Click the Optimization calculator button on the right-hand side. The Optimization calculator pop-up window will show up.

4. You can take a look at the data presented in the table and diagram depending on which value you are the most interested in:

Return of investment (ROI) radio button:

• Estimated ROI: Based on the input data, there is a 95% chance that return of investment will be within given range.

• Likelihood of being best: Likelihood of being the best performing entry in terms of ROI, based on the probability distribution presented in the diagram.

Earnings per visitor (EPV) radio button:

• Estimated EPV: Based on the input data, there is a 95% chance that earnings per visitor will be within given range.

• Likelihood of being best: Likelihood of being the best performing entry in terms of EPV, based on the probability distribution presented in the diagram.

• Estimated CVR: Based on the input data, there is a 95% chance that conversion rate will be within given range.

• Likelihood of being best: Likelihood of being the best performing entry in terms of CVR, based on the probability distribution presented in the diagram.

##### II. Optimize the Entities for Your Campaign Funnel

Once you have analyzed the statistical calculations, you might want to adapt that knowledge to your campaign funnel configuration. For example, you might decide to provide the weights for your offer to be able to distribute traffic diversely. Obviously, this is only a proposal how the Bayesian statistics might be interpreted and used to optimize the campaigns.

1. Go to the Campaigns  view. The Campaigns view will show up.

2. Find the campaign for which you have analyzed the Bayesian calculations and edit it.

3. Go to the DESTINATION option in the campaign configuration window:

• If there is only one path with one offer defined, you might want to create rule-based paths to distribute traffic between different offers based on their performance. For example, this allows you to send minimal traffic for the path with that browser rule where there is very low return of investment/earnings per visitor/conversion rate. Or increase the traffic distribution for browsers where there are many conversions.

• Once the campaign funnel has been split into more than one path, you can adjust the weights following the results provided into the Likelihood of being best column in the Optimization calculator window.

• Save the newly set campaign configuration and after some time check how the results have changed.