Thursday, March 26, 2015

Thousands helping Millions, Helping Your Federated Call Center Deployment

Thousands of volunteer experts have worked on Asterisk since 1999. This has made it into a powerhouse platform for telephony that you can use in your own system. Today it is a proven technology used by millions of users. Call centers and major PBX users are abandoning their legacy proprietary telephone switches to get the benefits offered by this Open-Source system.

Can you switch your large call center over to Asterisk? Definitely. The fact that Asterisk can run on commodity servers is a plus from a cost basis, but can introduce challenges in scaling. It is possible if you choose call center ACD software that can scale itself and manage multiple Asterisk servers. Each Asterisk server handles the media for some of your voice traffic. Over a number of systems, it can scale to handle a lot of voice traffic.

In a federated model, your deployment covers several geographic locations. This allows for voice traffic that begins and ends in different locations. Your agents may be distributed in multiple locations, but they can be active in the same ACD queues.

This would be expensive in the legacy model, but is an ideal use-case for Asterisk.

Federated deployment offers the advantage of effective use of skills-based routing in the call center ACD environment. Offering 24x7 service is easier if you have a geographically diverse pool of agents able to interact with your callers.

Tuesday, March 24, 2015

One Big Mistake You May Be Making With Your IVR

"The dialplan isn't working!" Four words that increase your stress without telling you anything useful. In the last couple of weeks, though, I've heard it a couple of times. Both times it was the same cause, so I thought I'd tell you about it.

You can't leave people hanging in an IVR (Interactive Voice Response). Silence is bad practice in writing dialplans, just like it's bad in radio. But, it's easy to miss some of the points where silence gets introduced.

Don't make his day any worse.
Web services: these always take time to finish. In your testing, they may finish instantly. In production, the network may get stressed. Your database might get overloaded. Your provider does something with DNS (Domain Name Service). Your web service takes 30 seconds, or worse, times out after a couple of minutes. Unless you've specifically set an audio file to play in the background or enabled music on hold, your caller hears silence. "The dialplan isn't working!"

Menu prompts: usually you'll background an audio file while waiting for input. Asterisk lets you pick a prompt, and have it repeat if the option selected is incorrect. Sometimes, though, you'll have a file that has to be played to the end then allow an option. If you don't set a sound file with your menu selection after that, though, the person may select an option they think is valid, then hear silence and an eventual hangup. "The dialplan isn't working!"

Other applications: another example is the administrative login for the MeetMe conferencing system. If the conference requires an admin to enter before it starts, and there's nobody else in the conference when the admin arrives, there is no audio prompt. In that case, the admin sits in silence. "The dialplan isn't working!"

It's vital that you examine your dialplan to look for opportunities for silence to enter your dialplan. Then you test and test some more. You can't wait for a network issue to delay your webservice and have all your callers abandon the call before they get through to an agent. Your dialplan builder should have the ability to add audio at the right spot, or do error checking, or play music in cases where there's no other audio. Take advantage of that.

Monday, March 16, 2015

Save Money, Do More With Asterisk-based Call Center Software

You can no longer doubt that Asterisk is here to stay in the telecommunications world. Over the last decade, the rate of adoption has been skyrocketing. Call centers are no longer complaining about Asterisk-based solutions, or worrying about having to have an Asterisk expert on-hand. With Asterisk providing a platform for contact center ACD (Automatic Call Distribution) solutions to sit on, you can transition from a legacy system to one built on readily available and mature software.

One of the benefits of Asterisk is that it provides a way to migrate your call center to IP telephony. When buying a proprietary ACD system, you have to realize that the underlying switching solution is a big part of the upfront cost. Asterisk gives you a telephony platform that has been deployed and used by millions of users worldwide. At a cost of free, Asterisk can give you immediate savings. When you consider its ability to run on free operating systems on commodity servers, the savings add up.

Proprietary switches and ACD software built on top of them are struggling to keep up. Asterisk can add, test, and deploy changes faster than the proprietary switches, and the gap is increasing. Any contact center software and next-generation ACD using Asterisk is going to give you that leg up in performance and features.

Once the telephony side is taken care of, business applications and CRM (Customer Relationship Management) systems can be developed and allowed to evolve. You need a system that gives you timely and actionable information that you can use to manage your client interactions. CTI (Computer Telephony Interface) allows you to integrate ACD and call control into your business systems. This ensures that business systems and CRM drive your telephony interactions with clients, not the other way around.

Indosoft's Q-Suite is a fully functional, feature rich call center system built on Asterisk. After a decade of experience with Asterisk, it can give you the dialer and ACD that your business deserves. A web-based agent interface comes with the system. External interfaces and libraries are available for you to integrate your own systems into the Q-Suite. This allows you to pull CTI functionality into your existing systems, or to create new applications.

Tuesday, March 10, 2015

Five Keys To Dialing Your Leads Effectively

Effectively queuing your outbound leads is key
The right queuing strategy is essential
You can have a great dialer, a wonderful staff, and effective management, but if your lead queuing strategy is wrong, your outbound campaign is going to fail. Great agents can't sell to people who aren't home. The best product for working couples isn't likely to be sold from 9-5. Your restaurant supply company isn't going to close any sales between 11am and 1pm local time, when the restaurants in question are struggling to handle the lunch crowd.

Outbound dialing has been on my mind lately, and so it's worthwhile to go over the factors you can use to order your leads when the dialer is queuing them up for your agents. Here is the list of factors most commonly used, in no particular order:

  1.  Priority: It might seem confusing to talk about priority as a factor in establishing priority. Its useful to have a condition that can be set as a tiebreaker for other factors, or to indicate an actual preference. It may be that everybody in Shelbyville loves cheese platters, but most people in Springfield have an aversion to cheese, so you may set it so that citizens of Shelbyville have a more immediate priority than citizens of Springfield.
  2. Newness: Call center staff, from the newest trainee to the CEO, love fresh leads. It can be a boost to morale during a tough day of dialing to let everyone know fresh leads are coming down the queue. Putting a timestamp on the lead at the time it was loaded into the system allows you to track several things, such as lead expiry. It also allows you to say dial the newest leads first. Old leads tend to be old leads because they've been dialed through already, but new leads have the magic of being untried.
  3. Disposition: there are a number of synonyms, but disposition is the one I'm using here. Leads that have never been dialed may have a blank or a preset disposition. Once the lead has been attempted or been handled by an agent, it will have a disposition. Some dispositions are better than others. "New Lead" is a good one, of course. "Busy" is better than some give it credit for. Sometimes a call center will reschedule Busy leads for tomorrow or further down the line, or even after "No Answer". That's probably a mistake. If the line is busy, there's somebody there right now. They just happen to be on the phone. The worst disposition is "Do Not Call". Don't call those.
  4. Time Zone: Time of day calling doesn't happen in a vacuum.  If you don't know the timezone the lead is located in (at least in theory), then you don't know if it's a good time to call. Calling east coast leads at 9am Eastern is a good idea for B2B dialing, or if you're targeting those who are probably at home. Calling west coast leads at 9am Eastern is almost always a bad idea, and may be in violation of the law. In North America, it's usually possible to guess a timezone rule based on the area code and local exchange, and databases of such are available.
  5. Heat: This is usually decided at the time of lead loading. "Hot" leads might be leads that were pulled from a web enquiry, an inbound call to another service, or collected at some event. The principle is that some leads are worth calling within as short a time as possible of collection. In the case of web leads, this could be seconds. You want to call when the client is thinking about you and anxious (or at least willing) to talk. Hot leads get queued for immediate dialing, with only hotter, newer leads getting priority. Normally, this heat factor expires after the first dial.
  Of all these factors, heat is the only one that should be a sole queuing factor, and it usually renders the others moot for the first dial. If you think about it for a minute or two, it becomes clear that the others aren't suitable for a sole dial criteria. If you decide that priority is the only factor for dialing, you wind up after a few days with all your high priority leads exhausted while you still had lower priority, brand new untouched leads in the database. If you call only leads in the Central European time zone, you may miss out on some great opportunities in France. If you only call based on newness, you're wasting older leads where the phone was busy at the time of the call. All too often, call center managers will point to one factor and insist it's the only one to consider, then change their minds once they've dialed their leads a few times. A default queuing mechanism that takes into consideration heat, disposition, date inserted and priority, in that order, allows you to call your newest "new" leads first, while still keeping busy, no answer and answering machine leads at hand for when the undialed leads have been exhausted.   

Wednesday, March 4, 2015

Post 100! It's Time To Celebrate

This is the 100th post on the blog, and my 40th. It seems like just yesterday that I started posting here, but when I look out at all the snow on the ground, I remember that August 20, 2014 was a long time ago.

As I noted earlier, the big shifts since the first post, which was a camera review, have been in the way VoIP has taken over, and the shift from the datacenter to the Cloud. In the next little while, the biggest changes will likely be in the widespread adoption of Unified Communications (UC) technologies and WebRTC. Both of these require a rethink of the interface for call center software and the processes in the center itself.

The next hundred posts should fly by a bit quicker, but I wouldn't be surprised to see just as much changed in the next 100 post interval.

Tuesday, March 3, 2015

Your Call Center by the Numbers, or How I Lost $2099.60 in the Stock Market

Understanding your call center software reported numbers is key
Oh no!
My wife says I should blog about the time I lost $2099.60 in the stock market. It was this week, so it's timely. She's one of my favourite people, so I'll indulge her. So, I logged into the website for my discount brokerage, and found that in the weeks since I last looked, one of my new favourite investments was down a whopping $2099.60.

I was a bit surprised. None of my "research" had indicated any issues with the company. Far from it. It's been doing well for years and had great prospects for the future. I looked around at a few financial sites to see what the issue was, and could only see positive reviews and great numbers. Maybe I had missed a raid on the head office? Maybe I had bought at a high (I did) and it had dropped down to more reasonable territory. I resolved not to panic until I figured out whether I had missed something big, or if I was just being subjected to the whims of the market.

This is the part of the post where I should get on with my actual subject, leaving you with a cliffhanger until you get to the last paragraph. I can't do that to you. Mainly because the resolution leads right into it. Within a few minutes I realized that the stock had undergone a 3-for-1 stock split. My broker can seem to get those right when it's Apple doing the splitting, but otherwise odd-numbered splits, or anything beyond 2-for-1, seem to wind up being wrong at least as often as they're done right. There's no real issue there other than a cosmetic (terror-inducing) error. They corrected my pre-split purchase price by a factor that was somewhere between  2  and e, rather than the expected 3 (or more likely 2). When I compared my actual purchase vs. my current holdings, I was actually up about $700.

When you're looking at the reports and wallboards provided by your call center software, you aren't likely to find the sort of discrepancy I found in my broker's reported numbers. However, sometimes you will find that there's a report that doesn't make sense in terms of the numbers. Let me use abandon rate as an example. It does come up from time to time.

Most call centers we encounter define time to abandon as the amount of time that has passed from when the call first hit the call center ACD system until the time that the client hangs up, assuming that they have hung up before their call is handled (usually by an agent). That seems simple enough. However, every so often, we find different definitions of abandons:

  • The amount of time that has passed since the call entered the queue until the caller hung up. I've seen a couple of clients use this measure, usually when their contract requires them to play lengthy greetings or pass the call through an IVR first.
  • The amount of time that has passed since the call hit the call center ACD until the caller hangs up, assuming it has reached a certain threshold. Your call center software may accommodate this as a minimum time to abandon. If somebody hangs up in the first 10 seconds, it's more likely that it was a wrong number than an actual interested party.
  • The amount of time that has passed since the call hit the call center ACD until the caller hangs up, assuming it has been in queue longer than 5 minutes. This is an extreme example, but some call centers take advantage of unclear definitions to use the above minimum abandon time to limit the number of abandons reported.
  • The amount of time that has passed since the call hit the call center ACD until the caller hangs up, unless it was transferred from another agent in which case it should only be counted from when it hit the queue. Ok...
  • The amount of time that has passed since the call hit the call center ACD until the caller hangs up, unless it timed out of one queue and then dropped into another queue. Well, ok, then I guess that makes sense. But what do we use for the abandon time for the second queue?
  • The amount of time that has passed since the call hit the call center ACD until the caller hangs up, unless it timed out of one queue and then dropped into another queue. Unless it came out of one of these related queues, in which case it is counted from the start of the call center arrival time. Seriously? So we'll need a way to specify queues that are related...
  • ... unless it passed through an unrelated queue first
  • ... or was transferred by an agent in this group
  • ... but not in this group

It can get confusing. Fortunately this can usually all be worked out while configuring the system. What does sometimes happen, however, is an admin will find a report that they hope returns the numbers they want, but that wasn't specifically designed for their own use case. This can usually be resolved pretty quickly in training, but sometimes does cause stress, unease and a little bit of panic if you take a few weeks to look at the numbers you need. Make sure of what you are looking at, and things will look a lot better.
Make your call center ACD work for you
That's more like it!

Wednesday, February 25, 2015

One hard worker among many: Team Spotlight

In one of the earliest posts on this blog, James introduced everyone to our brand new kettle, which he thought might inspire nightmares in some.

More than six years later, the kettle is still chugging along. It seems less nightmarish now. Given the number of tea drinkers in the office, it may be one of the hardest working members of the team. 

As you can see, it does have the looming spectre of the Keurig ever present over its shoulder. Knowing that its job is in jeopardy should it ever falter, we expect to see further prodigious output from this kettle for years to come.