Turning telesales success from art to a science

Selling products over the phone is never easy, but one Japanese outsourcer has used mathematics to triple sell-through rates.

Some say exceptional sales success is an art, but Japanese marketing consultancy transcosmos thinks it has turned the art into science by applying a range of data mining and textual analysis tools to refine the telephone sales process.

Based in Tokyo and Osaka, the company's 7,600 employees provide contact centre and marketing support to a range of customers in numerous countries. In this particular case, transcosmos was working with a Japanese health food company that was having trouble increasing sell-through rates to its more than one million customers. This struggle had kept staff morale low and frustrated efforts to expand the company's business.

Analysis by transcosmos revealed three major flaws in the company's marketing strategy: there were no clear rules for selecting the company's marketing targets; sales efficiency was compromised because so many calls were missing customers that weren't home; and the success of sales efforts varied widely between call centre operators, whose varying levels of skill made it difficult to develop unified marketing strategies. Overall, the ratio of sales successes to failures was just 2.4 percent.

"The idea of the client was quite fake," says Yuichi Ikusawa, acting division manager within the Service Planning Division of transcosmos's MCM Service Headquarters (MCM is the company's formal marketing service model, used to guide engagements with its customers).

"They just thought, with no logical proof, that their primary customers were housewives -- and they extracted the list of housewives from their customer database, and delivered them to the operators with sales targets. This meant some operators weren't using effective approaches to the clients with higher prospects, and the sales approach often fell behind schedule."

Sell smarter, not harder
Such problems are common in telesales operations, where delivering scripted sales pitches to large numbers of potential customers are critical for success in often highly competitive industries. Faced with the challenge of helping the customer improve its metrics, however, transcosmos took a step back and -- instead of subjecting operators to motivational speeches and endless retraining -- adopted a more high-tech approach.

A data mining-based analysis of the company's customer base, taking into account the customers' purchase history and demographic details, helped refine the list of sales targets. Rather than simply marketing to all housewives, the company now had a narrower target market that included women between 20 and 40, living in several specific areas of Japan. Based on the past buying history of these contacts, transcosmos set an expected calls-to-order ratio of 7.3 percent.

With the target market more clearly defined, transcosmos set about helping its customer figure out how to deliver the perfect sales pitch to its customers. Forty successful and 40 unsuccessful sales conversations were recorded and transcribed, then fed into a custom-built text mining engine that looked for common combinations of words, sequences of interactions and other patterns within the conversations.

The results of this analysis were plotted on a graph, which gave a pictorial representation of the arc of what Ikusawa calls -the ideal service conversation". With these parameters in place, individual conversations could be transcribed, analysed and compared against this arc, with calling scripts continually refined to bring them in line with the ideal.

Although the science seemed sound, Ikusawa says the company soon found that the approach met with mixed success: some contact centre operators were struggling to improve their performance, with motivation going down even more and the new scripts doing little to reverse the trend.

To resolve the problem, transcosmos launched another round of textual analysis, digging deeper into the structure of its sales calls and ultimately discovering that tendency to purchase was also correlated with other aspects of customer behaviour. The highest sell-through rates were found among targets who met three criteria: those who knew the client company, those who had an interest in health food, and those who had previously tried health food before.

This additional knowledge helped the company further refine its customer marketing strategy, comparing its list of prospective customers against the three criteria to reduce the sheer volume of calls the telesales staff had to make. In addition, creation of two subscripts helped the operator better adapt to changes in the flow of the conversation.

"This relieved operators from pressure," says Ikusawa, whose presentation on transcosmos's solution earned the company the Best Outbound Campaign mantle at the recent Asia-Pacific Contact Center World Awards. "They should not have to struggle. Through this collaboration, we can give the team the root causes of the gap and initial objectives; up-sell results were adequately identified so they could fix the problems in a timely way."

Selling better
Technological measures helped transcosmos identify the most common factors for its customer's telesales success, and to narrow down the target market to ensure telesales staff were spending time as profitably as possible. Combined with one-on-one reviews and involvement of contact centre supervisors, staff eventually began to show signs of improvement.

Within four weeks after transcosmos began conducting its text-mining exercises, the ratio of successful to unsuccessful contact attempts increased from 21.3 percent to 48.6 percent. This was a direct result of the narrowing of the customer's target market, which allowed the company to make educated guesses about the best time to reach each group of customers. With the hit rate increased, telesales staff were spending more time selling to customers, and less time leaving messages on answering machines or calling the same people over and over again.

Most importantly, the average order ratio increased from 2.4 percent to reach the company's target of 7.3 percent within just seven weeks. Compared with industry benchmarks of between three percent and four percent, Ikusawa says the company's telesales practice has gone from being a constant struggle to a shining star within the industry.

transcosmos has embodied its experiences during this engagement in its MO3 (Mathematical Operational Optimised Outbound System) methodology, which it now applies to the environments of customers in similar situations.

"The solution was not a quick fix, but an innovation in management," Ikusawa says. "Generally, the management cycle is incorporated in any contact centre, but in our case the management process was backed up with market analysis technologies that helped us bring sure and steady results."