Messaging Channels for Customer Service: Challenges and Practices
Big enterprises and major companies are already acknowledged of importance to implement non-voice channels for their customers — they are used to qualitative customer service, which is impossible without text channels. Connecting a Telegram or Viber channel, leading competitors get from 2000 to 4000 requests there in first month only.
What’s important to point out is that these results don’t require that much time and effort — adding a button on company’s website and simply making an announcement on social media is usually enough to drive attention.
However, every channel, text-based or not, has its own specifics and is designated for a certain audience segment — some prefer solving business questions by calling while others are comfortable with keyboard and messenger.
Businesses are already acknowledged — challenges still remain. How to organize a proper text-based communication with customers in company’s contact center, make it accessible for them and deliver best experience available while getting important stats and analytics on the case?
Improper understanding of these factors and lack of expertise in the field are main holding factors against implementing text-based customer communication channels. Transposing a generic call-based contact centre functionality onto a text-based one is among common mistakes too. Yes, they deal with similar questions, but their specifics differ drastically.
Common setbacks while implementing text-based channels are:
- Predicting how hard the channel will be loaded with inquiries
- Defining how many operators should work with channels and how to control their KPI and performance
- And a major one — implementing these channels without sacrificing the quality of service.
Gradually implementing text channels on website’s pages with lowest traffic in first instance, with more loaded ones eventually is among common practices in the field. This way it is ensured that channels won’t get overloaded with unpredicted traffic at first.
Yet this method has its setbacks — customer might not get his problem solved during first chat inquiry — operator tells him to contact call centre, forcing him to ask same questions again. In example, statistics tell that customer’s second inquiry on the same question to the logistics company lowers its income from his order twice, while third one lowers it down to zero.
Analysis
Upon deciding to introduce a text-based channel to your customers, ask yourself following:
- How many operators should be online at the same time?
- What channels and how many dialogues from them can they process simultaneously? (Consider this, as channel specifics differ greatly, if its a messenger or a website chat)
- What questions will operators process in every channel?
- Where will they look for answers and how long will it take for them?
Analyzing and answering mentioned above questions is important for a C-Suite to take a decision on whether to implement the text-based channels message processing. Not as easy as it might look.
Calculating voice-based operator’s efficiency metrics is simple. While knowing how many inquiries it can deal with, number of minutes for each, and a price of one minute of a call is enough, text-based channels require a deeper analysis.
Conversation lengths differ more in text channels rather than in voice-based ones. Operator’s answer time and number of conversations held at the same time depends on how many information is provided by its contact centre and how its work is organised. It is usually recommended to assign less experienced operators to 1–2 channels and from 3 to 4 for ones with some expertise.
Preparation
Before implementing text-based customer communication, company doesn’t know about questions to be asked there. That’s where the difference between spoken and written communication is seen — customer has a certain specific question in mind when he’s calling and wouldn’t mind to be held paused for a bit of time. Whereas chat-based communication allows customer to ask less specific but still important for him questions, i.e. loan interest rate or a schedule of some kind.
Still, there’s a risk of spam and irrelevant requests — they load operators hard and you want to avoid this, as it seriously affects their efficiency. Following might help out:
- Understand what kind of inquiries will be received in each channel and solve them efficiently.
If the channel’s operator is raw in a customer’s particular question, it might be better to switch chat to a more fit operator rather than forcing the customer to access another channel.
It is better to automate the process — depending on the page the customer is on, a basic set of his questions can be predicted, which routes him to a needed operator. This can be applied with chatbots too.
2. Know the peak loads: when and how hard the inquiries are going to hit?
WFM systems are in use to automate staff time management — it receives traffic and analyses the amount and time of loads. It also gives data about what kind of traffic goes through the channel, how hard it is now and predicts its further amount. Also, WFM systems give prognosis on how many operators will be required to handle these loads.
Mentioned above makes them a useful tool to lessen the contact centre staff headache and improve its efficiency.
3. Set the Key Performance Indicators.
These efficiency metrics, KPI, are in help for department heads and top managers to keep track of contact centre staff’s performance and arrange motivational models. This method applies for text-based channels same as for voice-based; analyzing a set of listed below parameters might get you a lot of information:
- TST — operator response time; it is usually set from 10 to 20 seconds.
- ACD — average conversation duration; average indicator is 5 to 7 minutes and 1 to 3 for messengers.
- LCR — lost conversations rate; customers that haven’t made it till conversation and left the dialogue before operator had answered. LCR is a high level of importance indicator, as customers that closed the chat window can’t be contacted back again. Acceptable rate for major contact centre is 1% and less.
- CSI — customer satisfaction indicator; 85% of positively rated dialogues is considered as a good rate.
Surely, set of KPI is specific for every situation, however, these four major ones are the ones that are best not to be ignored.
4. Set up template responses and knowledge base.
Knowledge base is a company’s all-department unified informational system, accessible through table sheets, documents or embedded into comfortable interface. It can be used as a help with particular product for a customer, co-worker or in means of self-education. If compiled properly, a chatbot can be connected to it to respond to customers on first support line.
Search efficiency in knowledge base is also a very important stat, as it directly influences customer messages processing speed, which in its turn influences the project’s overall end profits, so the more convenient and structured your knowledge base is, the more operators are efficient.
Subtleties and pitfalls
Chat is not always a salvation — sometimes there are topics not to be discussed with text.
For example, banking that subdues to normative legal acts, regulating the terms of work and service of people and organizations, defining what information and in what cases should be requested and transmitted.
Restrictions also apply to chats in authorized areas, and what can be discussed on voice channels cannot be written, for example, asking the customer to write his codeword. Properly compiling sets of allowed topics for every channel — phone, website chat and messenger will allow to start designating operators for them.
There are situations when channel operators can work both in-state and outsourced. It is usual that outsourced contact centres work as a reference service, while more detailed and specific questions and queries are handed out to in-house operators team. Maintaining the work of in-house contact centre might be considered expensive, but it drives the service quality to high levels.
With both contact centres connected and interacting, it is possible to direct and separate the incoming queries into channels:
- Chatbot — automatically responds to most common questions using a set of template answers;
- Outsourced contact centre — provides customer with reference services;
- In-house contact centre — provides customer with specific and personalized information, if required.
Simply incorporating a chatbot and leaving it be is not always a good option. As it should be supervised in separate window, this only adds more problems with multitasking instead of serving as a solution — average call centre operator’s software arsenal consists of up to 8 different applications.
Coordinating operators between text and voice channels is also not as easy as might seem. Pitfalls are caused by software specifics — switching channels from voice to chat and vice versa causes misconductions in stats — closing chat in order to switch to another interface indicates that operator has finished his shift.
Gathering statistics from voice channels is a common practice for contact centres, and receiving them from text channels as well might be not as easy — only if chat channels are integrated with voice platforms. Integrating these statistics is also a frequent problem for contact centre managers — keeping track of them separately is close to impossible, and hiring a separate employee for this task only doesn’t seem like best option available.
The only available solution to all of the mentioned above seems to be a unification of inquiries from every channel into a single line and data from them into a unified statistics system. Platforms with described functionality already exist, however be sure to clarify following moments with service provider:
- Data storage: how it’s realized, how long and where it is stored, and how the integrity and confidentiality are provided?
- Data transfer: know the encryption protocols.
- What information can be processed by vendor?
Summing up, proper incorporation of text channels for customer support contact centre lies on two major pillars:
- Organizing inner rules and regulations concerning areas of consultancy — what information and where shouldn’t be mentioned.
- Setting up contact centre’s inner processes — platform setup, data processing, supervision, staff training and coordination.
Successful result here is not an easy task, and achieving one is available by incorporating an ready-made omnichannel solution, such as BRN.ai offers, either having a professional with this kind of expertise hired.
Start your own Chatbot Company with a White Label Version of BRAIN
Alex Galert CEO BRAIN (brn.ai)