Okay, so no-one around the BBQ is excited enough about IVRs to sing their praises. Most the IVR developers I know are to scared to say that they responsible for these things, but it’s time this situation is clarified.
Firstly there are many IVRs that serve customers needs and even score higher satisfaction than a human undertaking the same enquiry. In fact, best practice open speech systems are now almost seamless. Take Westpac as an example where 80%+ of callers are exited at their first request. No repeating yourself or confirming what you said. One and done. i.e. say it once and you are finished, on your way to the right service.
But probably being more critical of the situation, when you think of IVR v concierge, the COMPARISON SIMPLY ISNT RELEVANT
What the researchers need to understand is that it’s actually about whether a customer wants to be served or offered a reception. If we assume that a business isn’t going to add to its service costs (or these would be passed on anyway to the customer and we’d need to add that into the mix) then by offering a reception we are in essence removing resource from actually dealing with the enquiry. Now ask the customer if they would rather navigate a few steps in an IVR and then be served or get a reception but no-one available to deal with their enquiry.
What’s worse is that the reception model is also a bottleneck to the organisation. It may look okay until normal demand but it will snowball under peak demand. We have run side-by-side speech router versus human concierge and discovered:
- a 5% uplift in satisfaction with the human concierge (under normal demand) but a slight decrease during busy demand times (and by the way, that’s compared to a speech router that lacked some of the best practice learnings now evident in recent deployments like Westpac, Inland Revenue, and Lloyds bank where customer satisfaction is much higher)
- increased routing error and variability with the human concierge (typically the reception isn’t expert in the business structure) often leading to subsequent transfer problems
- significantly more routing end point knowledge with the automated system (dozens of specific route points understood versus a handful by the human concierge)
- loss of self-service opportunity when directly routed to human assistance
The concierge challenges are exacerbated by customers expectation that they expect the first person they reach to be able to deal with at least simple enquiries. They become agitated when they have explained their problem and even when its perceived as simple, they require a transfer.
The problem isn’t straight forward with complex enquiries either, as determining this is the challenge. If the customer invests a lot of time in explaining their problem they will end up repeating it again when they are transferred. However, if they over simplify it, say by just describing the product, they are likely then to need to be transferred at the next point of routing!
So then consider the ‘pebbles and boulders’ approach. That is where the first tier agent does attempt simple enquiries. The result: well, the simple, low value callers are happy but those with more challenging problems, and anxiety as a consequence, are destined to be transferred around the business again. In other words that model delivers good call answering capacity, something that appeals to contact centre management because they are measured on service level (answering calls inside time perimeters) but not for the most valuable, most at risk contacts. Our experience with these models is they generate around 35% 1st time transfers and then up to 10% multiple transfers, because the 1st tier agent isn’t as knowledgeable about where complex calls should go.
Compare this to an open speech IVR where customers can request things in their own words. While it isn’t perfect the result WILL be matched to the agent best able to deal with that enquiry. No-one is tied up bouncing calls about or on 3 way calls, releasing the potential to get on a deal with customers calls.
My Vote: I’d prefer to go directly to the right expert but I’d rather press a few keys on my phone than navigate a series of people transferring me about to get that right service. Put simply, don’t waste my time!
And for those wanting tips on how to best navigate an open speech system: Use normal terms, and be specific. They’re not all great today (yes some are terrible) but best practice is changing that, so you’ll increasingly see that these systems do work in a natural and value adding way.
By the way, don’t: yell at it, over emphasise your words, try and bypass it (you’ll just end up somewhere wrong)
About Flare Design: Flare are leaders in open speech customer experience and business process design. We bring a complete end-to-end view of the impacts, pitfalls and mitigating strategies of open speech implementations, change and the impact on staff. In the last 10 years, we have been fortunate to work on many of the major speech implementations and have contributed to the significant turn around in customer perception and press adoption of speech recognition systems.