OK,

Your starter for 10 points.

What is the minimum number of OB agents you would
THINK you could use a predictive dialler with?

I heard a number yesterday I couldn't (and still can't) believe.

DaveA

If it was network based then however many agents you can log onto your network...

Regards,

Ian

Of course it would help if I actually read the post properly... :-)

Most predictive diallers actually operate in progressive mode until 8 or so agents log on, then after a short stats gathering period they will switch to predictive.

What number did you hear?

Regards,

Ian

4 on a predictive :-)

DaveA

It all depends on the dialer, an Avaya PDS dailers is recommended to operate on a minimum of 50 agents for maximum return on investment.

I have heard of people using PDS with as little as 5 agents. The trouble is the costs involved in implementing a dailing solution. An Avaya PDS is widely regarded as one of the best dialers in the would and obviously has a matching price tag so if your running 5 agents on it when the ROI was based on 50 then your not getting value for money as the dialer is not working to its full ability.

You can look at a solution like Altitude which is recommended to be running upwards of 10 agents and is a soft dailer.

Soft dailers have lower answerphone detection rates, lower call connection ratios and are (sweeping generalisation) not as good as hard dialers. They are of couse not aimed at the same market space though.

The hard dialers are the workhorses of the dialer world and are aimed at 50+ simple as that.

Jason

Note the very slight case of dyslexia.

I have to disagree with Jason's sweeping generalisation that hard diallers are better than soft diallers. I can point to experienced users that prefer their soft dialler to their hard diallers as they get better performance.


Regarding the minimum number of users on a predictive dialler... it all depends on the length of calls, the quality of the dialling lists and how the dialling algorithm works. As stated by earlier respondents, the number can be very small, but the dialler will throttle back to 'progressive' mode to ensure that nuisance calls will not be generated.

It should also be recognised that if the generated calls are fairly long and the contact rate from the calling list is reasonably good, then overall productivity for a 'progressive' dialling is close to 'predictive' performance.

The thing to remember is that a dialler - even dialing at 1:1 is still far more productive than a manual system.

I can only agree with Ken, it all depends to such an extent on the quality of the data and penetration - the better the data, the more agents needed to benefit from predictive dialling. With a list with very low contact rates - then 4 agents could benefit - though I think that the number is low.

Bear in mind that the benefit is exponential - doubling the number of agents more than doubles the benefit.

Richard.

Hi Dave,

I am relatively new to Diallers and this site. This has been one issue that comes up daily and what I have seen is that for the agents to get calls at a good pace 10+ agents are the minimum required though the dialler would even run with 4-5 agents but would not be that effectively.

Hope I got that correct.

Let’s try and give some guidelines to the questions in this thread

1. How many Users

This depends on..

• the quality of the predictive dialing solution
• the amount charged for it
• the kind of campaigns being run.

In practice these conditions can be so variable that buyers simply have to be aware and do their homework. Best is to get a free trial. And beware the reference visit where you have no idea how to relate the dialing and campaign conditions you see to your own situation.

The chap who suggests that you need 50 agents to get decent benefits is smoking dope. Nuff said.

Any dialer really worth its salt should start to provide measurable benefits at around the 4/5 agent level, dialing under compliant conditions. It’s a baseline we usually set for our dialer, but then it is subject to the customer ROI, and depending on the payback period sought some users might look for a higher number. Remember this is agents on a campaign. Users running lots of small campaigns might want to follow inbound practice and combine campaigns, with agents able to handle multiple scripts. Gets agent numbers and performance up, and makes the agent’s job more interesting. As more users seek to make their dialers work under compliance, expect this to happen a lot.

2. Soft v. Hard Dialers

It should surprise nobody that the soft guys think they are best, and likewise the hard guys. Our company does both, since it’s horses for courses. In a European world, with an efficient switch, end to end ISDN, and answering machine detection sensibly switched off, there isn’t a fat lot of difference in performance, and this is the market that most users probably inhabit.

But if you are located in the US, with the lack of end to end ISDN coverage and the persistent belief that answering machine detection is the way to go, you might be more inclined to a hard dialer.

3. Other Points

Doubling the number of agents more than doubles the benefit?? In some cases maybe, but this is not a general rule. In a well-engineered dialer operating under compliance, then it might be true if you go from 5 to 10 agents. But it is very unlikely to be true if you go from say 25 to 50 agents on a telemarketing campaign. The incremental benefit per agent will tail off. Remember there are only 60 minutes in the hour.

1:1 trunk ratio. This applies to manual, preview and progressive. You can certainly improve performance by automating matters as you move up from manual, but you will be lucky to get much beyond 30-35 minutes talk time in the hour, unless you are doing something like market research. Beyond this, predictive dialing, with an increased trunk ratio beyond 1:1, has historically held out the promise of 50+ minutes talk time in the hour on a campaign, but operating under compliance in today’s markets, this a challenge too far many times, for even the best dialer.

Ian Robinson,

I seen most of your posts and I would like to ask you first before I go and call Concerto. Would you happen to know if you can use Crystal Reports 7 with a Conceto Unison Dialer? And if so would I just setup an ODBC connection to the dialer? The reason why I ask is that my company is switching from an old Melita dialer to the new Unison dialer. I do know the dialer software comes with its own reporting package called Center View, but how our VB programs are setup we use Crystal and Access for about everything.

Thanks for your help,

Stephen

Stephen, I don't know the answer to your specific question but should you find that the Unison doesn't meet your needs then I can advise you that the rostrvm dialler has easy VB integration, ODBC integration and off-the-shelf Crystal Report templates (click our ad on the General forum to learn more)

[end of blatant plug!]

Stephen

Yes you can use your Crystal reports 7 with the Unison dialer. You can set up an ODBC connection for access. Contact the Concerto support desk if you need any help or let me know if I can be of assistance.
Tiffany

Hi Stephen, as Tiffany has stated you can report on Unison with Crystal (or Brio and other related)..

Another plus is easy integration with ACD's, desktop databases etc, pull all your call handling platforms into the same suite of reports..

If only Concerto would release a database schema to the general public, then again the reporting package that you can buy from Concerto would look a lot less attractive (and who can blame them really!).

Hope it helps,

Regards,

Ian.

There is a very good schema of the ComposeIT database (Programmers guide to the ComposeIT database). It is a complex document - but certainly gives all the links, keys etc.

To a report designer is is the mother lode. Many customers use it and swear by it.

Richard

I use the CRS Mercury predictive dialer here for all my campaigns and it has a variety of settings from 'don't dial if no agents are free to take a call' to very agressively predictive. In the case of one of the agent groups which is very small (4 agents) there are times when the dialer is running in (low) predictive mode for 2 or even 1 agents (due to break schedules) without excessive abandoned contacts.

I wouldn't say that this is a very effective use of a predictive dialer, but it is possible.

RR:- "To a report designer is is the mother lode. Many customers use it and swear by it."
..or swear at it when they can't get it working :)
It is very comprehensive and you can get just about everything out of it that you'd need. I've been using ComposeIT for a while and when you suss it out, it is an incredible piece of kit.

Any other diallers I've done reporting on it's been SQL data backend and crystal or Excel/ Access reporting. Which is more "homely" because you generally have a great skill set for Excell in an office environment than anything else.

Minimum number of agents is a tricky one, I totally agree with Michael. Even if you have 30+ agents on a dialler, if you data is bad then you'll just burn through it or over pace to compensate and drop when you do connect.

What we need to run dialler effectivly is:
More agents
Better data and
Good agent scheduling.....oh for an ideal world.

I think with Crystal, as long as you have a data source or a direct ODBC connection you can report on anything.

Cheers,
C.

Alright Wallsy, where you working now, I thought you moved into the conferencing business? Sorry I know this is irrelevant to the post

To add a slightly more relevant post, in general 6-8 agents are recommended for predictive dialling does depend on external factors however in terms of Crystal Reports, if the diallers backend is one of the mainstram RDMS then you can use it ORACLE/SYCASE/SQL Server/DB2, it may also be possible to download ODBC drivers if you have a more obscure backend. Wallsy no longer at excel just remebered my old login

Mr Tonner,
Again apologees for the irrelevance of this post.
Jeff drop me an email to my hotmail account.
"Colin_walls"

Cheers
C.