Guest Post: How Teletext Holidays has focused on non-bookers to improve brand sentiment

Guest Post: How Teletext Holidays has focused on non-bookers to improve brand sentiment

Steve Endacott, chairman of Teletext Holidays and sister technology firm Zen3, explains how call centre analytics developed to use artificial intelligence has been deployed to boost how the brand is viewed by callers who don’t convert Continue reading

Steve Endacott, chairman of Teletext Holidays and sister technology firm Zen3, explains how call centre analytics developed to use artificial intelligence has been deployed to boost how the brand is viewed by callers who don’t convert

I must confess to being a brand convert, believing nothing is more important than building brand traffic to reduce ever-increasing Google advertising cost.

So it’s nice to be chairman of a brand like Teletext Holidays, which I’ve interacted with for over 25 years now.

Teletext is lucky to enjoy unprompted brand recognition levels of 40%, which puts it on a par with UK travel giants such as Thomson, Thomas Cook and Expedia.

This alone delivers significant volumes of direct brand traffic via SEO and PPC channels, reducing average customer acquisition costs compared to other OTA players such as Travel Republic, On the Beach and Love Holidays.

Unfortunately, our customer consideration is much lower as many lost touch with Teletext when it ceased to power the pages behind their TV and moved fully online.

We are addressing this with above the line TV advertising and sponsorship deals such as the recently announced tie up with newly promoted Sheffield United (Another historic giant on the rise!).

Legacy brands such as Teletext Holidays also offer a key advantage that customers already know what the brand stands for.

Extensive research shows that customers view Teletext as a late deal offer site for beach holidays booked by phone.

To offer the by phone service we outsource call fulfilment to Truly Travel’s Indian based call centre, at a considerably reduced cost versus an equivalent sized UK call centre.

Conversion levels are extremely healthy compared to any UK call centre, and our ability to switch and directionally sell products means we have enjoyed strong margin growth.

However the Net Promoter Scores were not at an acceptable level when I joined and have been a key performance indicator that we have been working hard to improve.

It’s my belief that in today’s world of social media and review scores, focusing on the customers that do not book with you has never been more important as word of mouth and review scores can kill or make a brand.

However, like many call centres, our metrics and focus was primarily on conversion levels and profitability per call alone.

In order to find a cost effective solution to this problem we looked towards our sister technology business Zen3, and worked with them on the development of their Sayint speech to text system which has revolutionised our approach.

Sayint allows us to record every inbound and outbound call into the call centre and then translate these into written words, so that they can be data mined using the latest big data Artificial Intelligence algorithms.

Sayint has allowed us to create sentiment algorithms weighting basic factors such as call length, hold times, silences etc and then overlay them with scoring based on the presence of both positive and negative phrases.

Some examples of negatives are phrases such as “Can you repeat that, pardon, that’s more expensive, I don’t want that” and phrases like “Can I talk to your manager”.

Over a period of time we have built algorithms that we use to automatically rate a call, in terms of customers satisfaction levels.

These allow us to generate a ranking by agent and an understanding of which agents are scoring well for service whether the customer books or not.

The correlation between top seller and generator of highest customer satisfaction over all is often not what you may expect.

The tool also allows managers to walk through calls with an agent, drilling down in the written conversational record to where that negative phrase occurred, so they can then listen to that exact section of the call and coach a better approach.

This has allowed us to focus management review and training precisely where it was needed, which in turn has increased the average satisfaction levels on non-booking calls by 26%, as well as increasing call centre conversion by 15%.

It is, however, the improvement in satisfaction levels on non-booking calls that Teletext continues to focus on because this is both where its ability to increase profits lies and the biggest numerical influence on its average review scores in sites like Trust Pilot and Feefo.

Good scores in these areas, in my opinion, make customers more likely to click on your brand adverts when they see them or book with you if they are looking for that third party reassurance.

Sayint also has allowed Truly to reduce its call audit team from 10 to 4 whilst increasing the volume of audited calls by fourfold, by being able to accelerate the speed at which keywords in booking calls can be found.

Like most large call centres, I know we have and continue to have quality issues to deal with due to a relatively high staff turnover, but at least management now have a tool that gives them the measurability and visibility to force the required action to make improvements.