A bit of magic: how to take and make a call-center really effective

You call provider. When you are ready to talk with a tortured and cheerful girl about the number of green light bulbs on a black box, you even get a bit lost when a natural sysadmin answers you. And immediately understands the problem and solves it. You hang up after 25 seconds of conversation in light shock.

Then you call back from your sister's phone and get affectionate “What color is your Internet?”, And you begin to understand what the matter is.

Yes, it is actually possible in practice. For example, let's take some typical call-center and delve into its already collected data, and then connect some mathematics.
')

Mammoth peers

Usually, call-centers have problems with staff turnover, transfer curves between employees, poor service and other features, and efficiency is measured by call time and the number of processed calls. And this approach is outdated for a long, long time.

What are the input?

Firstly , we know about customers many times more than it seems at first glance. Immediately after the call, we can divide customers into groups (VIP, beginners, regular customers, outsiders, random calls and more), which means a different attitude.

Secondly , oddly enough, the call-center is the same selling unit as the sales department. The first consequence is that more professional specialists respond to calls from more interesting customers after segmentation. If the call is random, it can be transferred to the intern. If the call is from a client with a two-year history, it must be processed by a specialist who will not be mistaken and will show the same high level to which the client is accustomed. The hierarchy of sellers-operators is one of the motivational factors, which reduces staff turnover and increases involvement in the process.

Thirdly , in modern call-centers people are frightened by the following factors:

• Clear and specific responsibilities;
• The strictest discipline;
• Lack of initiative;
• Work under constant control;
• The monotonous and monotonous nature of the work;
• The lack of a creative component in the work;
• Arrhythmic loads;
• Presence of non-standard situations.

It is clear that it will not be possible to defeat everything at once, but it is quite simple to remove 90% of problems with the help of the tools that Data Mining gives.

What to do?

For starters, collect some more data. Example: our call-center in addition to transactional data adds some data about the interaction with the client and the emotional colors of the conversation. All the data that we already have or will have, we can process and with the help of various systems get a forecast of care, or segmentation, or something else. At the same time there is a rigid mathematics, it is described in the theory.

For example, cluster analysis - identification of clients with similar behavior, associative analysis shows interchangeable products and services, classification and regression analysis make it possible to predict the response of clients to an offer.

How to use it?

Turn operators into sellers , but not just saying “now you are a seller and are responsible for our key customers”, but by giving the appropriate tools. For example, if there is an understanding of which segment a particular client belongs to, then you can transfer it to a specific operator. This is not needed by the seller, but it is necessary so that the incoming or outgoing stream of calls and letters is somehow filtered.

Here comes the bell. Operator-seller already sees the profile of the client and understands what time he should spend on working with this treatment. From where It's very simple: the data is formed individually: someone likes to talk longer, someone on the contrary needs to be quick. If there is a predicted time, then the operator knows that this client needs to quickly and clearly state everything and end the conversation, while with the other one can talk longer. Imagine an idyll when the first line of your operator already sees your profile on the screen, which says that you reboot, turn on and off, look at the lights and close-open BEFORE the call?

And to this our call there is no KPI "under the comb." Parameters are selected for the client. For him, this is an excellent service, for you - the opportunity to motivate the staff and additional profits, because the operator can easily make an additional sale to a regular customer without getting out of the KPI - and all parties will be happy.

Four breaks! Four!

We look further. We can collect the emotional characteristics of customers. A rubricator can be made: as far as the client is talkative, how loyal he is to us, the operator can simply tick off the appropriate boxes during a conversation with the client, focusing on his feelings. All this greatly affects the conversation. Plus, we see non-standard situations: for example, screaming and nervous customers call any call-center. But when the operator puts a corresponding tick, he sees an alert that the client of the previous 10 calls has always been calm and collected - and, therefore, he was hooked on something vital. And we need to urgently solve this, for example, by connecting an additional service or by doing something else outside the framework of standard practice. And the call-center software can open such powers to the operator, plus show what options there are.

There is a tool to enable us to offer something to the client. For example, the propensity to care indicator gives you an understanding of whether to retain the client or not. This is a question of how much we are willing to spend on leaving the client “ours” - and it is automated and given to the operator. Marketers are now, understanding what is happening with this, probably already experiencing something like an orgasm.

Now outgoing calls and letters. The operator can determine the sequence in which to call. First, we work with the most profitable or those who tend to care more than others.

What to do with trainees? To give them the most uninteresting financial clients, but take into account the number of “good” transactions, and then increase them when a certain result is achieved.

Not convinced? Then some statistics:

That's why it's worth it:

• 75% of customers prefer a high level of customer service, while 15% of them are willing to pay more;
• 50% of customers go to competitors due to poor quality of service;
• 20% of customers tend to go to a competitor because of dissatisfaction with the last communication;
• 86% of customers welcome the availability of personalized offers during self-service operations;
• 84% of customers are willing to pay more for better service quality, of which 56% of customers are willing to pay 10% more, 24% of customers are willing to pay 15% more, 20% of customers are willing to pay 20% more.

Summary

When we have the tools that we receive thanks to technology, we can say that we can divide clients, know how to work with them individually, and we understand that employees need to be divided into different categories.

In my practice, the right Data Mining frees up the huge hidden possibilities of a call center. This is not a panacea, but the tool is really almost magical. It is like reading thoughts, only this is predicted, calculated and turned into business tools.