Ringba's predictive routing feature chooses which target to send a call by calculating each target's estimated revenue per call (RPC) and sending the call to the most lucrative target. You can find more information, including step-by-step instructions, in the Predictive Routing article.
You configure the logic predictive routing uses to calculate RPC when you create a predictive routing profile. If you use the advanced tools, you specify the Minimum Number of Calls and Desired Number of Calls you want the system to evaluate for each target when calculating the RPC. You can learn how these settings work together, and about other settings that give you even more flexibility and control, in the Predictive Routing - Advanced Use Cases article.
The values you choose for the Minimum Number of Calls and Desired Number of Calls fields tell the RPC calculation how much of a target's most recent performance to consider, so they influence how quickly the RPC calculation responds to changes in a target's productivity. There is no "magic number" for either of these fields get the best RPC calculation. The best answer for you depends on your call volume and your business priorities, while also considering your vertical, the partnerships you have with your targets, and how aggressive you want to be.
Call volume and time
Consider these examples:
- Account A runs an average of 500 calls per day per campaign and target.
- Account B runs and average of 250 calls per day per campaign and target.
Both accounts set their Minimum Number of Calls field to 250. That means Account A considers the most recent half-day's worth of calls when calculating RPC, while Account B considers a full-day's worth of calls.
If a target has a sudden decrease in performance, Account A's algorithm notices it within half a day and can adjust to sending calls to higher performing targets. Account B's algorithm won't notice until the next day. The same can be said if a target has a sudden increase in performance: Account A's algorithm can send more calls to the target on a hot streak faster than Account B.
Notice how both accounts set their number of calls to the same number, but the results vary based on the call volume.
Stability versus responsiveness
The above example might sound like quicker responsiveness is always better, but it depends on your targets. You know your targets and your relationships with them. A target can have a string of very good luck or very bad luck: whether you want to make quick changes to your call routing based on short-term changes to conversion rate also depends on how you work with your targets. For some people, stability in call routing is also valuable.
Bumps for new and underperforming targets
The number of calls you configure the algorithm to consider is not the only factor in determining which targets get calls. You can also configure priority "bumps" for new targets that haven't had enough call volume to make a fair comparison and temporarily underperforming targets that need a chance to catch up.
Find what's right for you
Fortunately, you can adjust the settings you configure in your predictive routing on an ongoing basis. As volume changes and your relationships with your targets mature, you can tweak the values to maximize your returns. You should make adjustments slowly and observe the results for a few days before making more to give enough time to understand the impact of your changes.
For more information on all the advanced tools you can use to configure your predictive routing, including step-by-step procedures, see the Predictive Routing - Advanced Use Cases article