Hello

I'm looking for some information on the perfomance increase on a predictive dialler depending on how many people are logged on.

What I'm trying to demonstrate is that for example the performance increase of having 10 people on is more than double that of having 5 on.

Has anyone had to demonstrate this? or got any suggestions how this could be done?

Thanks

In my experience most dialer vendors recommend you have at least 8-10 people on the dialer as it allows it enough data to properly predict call distribution. It's harder with less people as one person gets up to walk over to the fax....and that's 20% of your staff if you only have 5 people.

It's tough to get your dialer to properly queue up calls with those kind of flucuations.

I don't think once you get past 10 people or so you would see much difference though.

I am not a dialer expert but I know we have a couple on here.

Generally it isn't worth using a predictive dialler with just 5 people, unless your calls are very high-value (in which case, abandoned calls probably aren't good for your brand and you need to hire more people instead). However, using a decent progressive dialler can still make a significant difference to productivity and success rates.

My experience is that you only get around 5 minutes talk time/hour improvement with a predictive dialler over a progressive dialler with 5 users. You start getting good productivity at 10+ users, and it improves as the number of users increases.

Once you have more than around 15-20 people, you get additional improvements - agents start coming off calls more often, so the number of potentially abandoned calls that get "rescued" increases, which means that the dialler can take more risks with the quota of abandoned calls, which means it can make more calls. This is why Answer-machine detection can actually reduce performance.

Connect rate and average call duration also make a huge difference to how effective a predictive dialler is - ideally you want a low connect rate (but not too low!) and a short average call duration.

Hi Dan,

I hate to answer a question with a question, but what brand predictive dialler are you using?

I ask because, yes it is true that the number of agents logged into a predictive dialler will have an impact on productivity, if by productivity you want to decrease wait times between connects. However this does depend of which dialler you use.

Some dialler manufacturers will quote a minimum of 4 agents to run an effect predictive dial, some up to 10. In all honesty the predictive dialling algorithm is much more complicated than to make a significant effect with agent numbers alone. I have consulted at some Contact Centres which run a predicitve dial with 25 agents that do not get anywhere near the same productivity as a Centre with 5.

Every dialler vendor will have their own test results for effective predicitve dial and these are what I always use a minimum requirement.

When it comes to demonstrating greater productivity with more agents - I would much rather look at maximising productivity with your current agent resource until this is at a point where greater agent numbers is the ONLY way to improve contact rates, conversions etc.

Doubling agent numbers will not double productivity persay - if only it where that easy.

Instead significant increases in productivity can be made by looking at the following

1. Average number of connects
2. No of lines:agent ratio prompted by ave number of connects
3. Use of Answer Machine Detection and its relative accuracy (test it)
4. Drop rate settings
5. DATA MANAGEMENT - So important - productivity is directly related to how you manage your data set, load them and churn them.
6. Agent activity - should aim to be consistent and somewhat predictable i.e. waiting, talking, wrap for majority of the day with very infrequent pauses and logging off.

In closing a predictive dialler is using an algorithm to calculate availablility of agent resource. It's taking into account group and individual averages. Therefore the bigger the pool the more accurate the average calculated and the less likely a drop or an increased wait time, but by definition the better the agents time management and consistency of call the less important the number of agents.

Thanks