Richard Snow looks at what is wrong with IVR systems and what we can do about it.
Interactive Voice Response (IVR) was introduced into contact centres for three primary reasons:
1. To collect customer information before putting the call through to an agent
2. To deflect calls from agents by handling the customer’s query within the IVR, e.g. allowing customers to ascertain the balance of their account.
3. To provide customers with information in the hope they would decide they didn’t need to speak with an agent, e.g. if you are calling to find out why your phone is not working, it is because the network is down.
When you get behind the reasons why companies say they introduced IVR then the bottom line is cost reduction – shorter calls, fewer calls, fewer calls needing an agent – all leading to lower costs.
Did it work? Well, based on research I did last year, yes, but only to a limited extent. Of the 253 responses that qualified as either managers of contact centres or business leaders, nearly 60 per cent indicated they had deployed IVR for one or more of the above reasons. But success rates were markedly lower, with 59 per cent indicating that in over half of cases the customer ended up speaking with an agent anyway.
Why IVR goes wrong
The foremost reason is that the scripts and menus are just too complex. Customers get confused, fed-up and sometimes downright angry as they struggle to pick the option that best meets their needs, which sometimes doesn’t even exist. Worst of all is getting stuck in a loop with no way out. Things have got so bad that in the US there is a website that for some of the more popular sites identifies what selections to make to get to an agent as quickly as possible!!
The next thing customers hate is when, having gone to the trouble of entering information to identify themselves, when they get through to an agent, the agent asks them the same information all over again – it is just like the IVR was never there.
The third most popular complaint is that the information provided is not consistent with information provided through another channel, e.g. the balance doesn’t match the one shown on the latest statement, or the information just invites the customer to request to speak with an agent anyway, e.g. the balance is not what they expected so they request to speak to an agent to find out why.
What can be done to improve IVR
In the majority of cases the answer seems to be “nothing”. Despite obvious levels of customer dissatisfaction, many companies opt to leave things as they are, thus not only failing to meet the original objective of saving money but also upsetting customers in the process. One of the most worrying insights from my research is that 11 per cent of the qualified respondents didn’t even know how many customers opt out of using the IVR to speak to an agent.
Today there is no excuse for companies not knowing how well their IVR is meeting customer expectations. There is software available that can collect the data about what options customers choose and map this back to process-like maps that show exactly what paths customers took through the IVR and when (and why) they opted out. Using these maps, companies can identify changes to menu structures and related information access to better meet customer expectations and their business objectives. For those that can’t be bothered, then in many cases the answer should be to save the money spent on IVR and put customers straight through to an agent. Or, more simply put, get it right or get rid of it.
Richard J Snow is VP and Research Director at Ventana Research (www.ventanaresearch.com)