by Liz Wanless, Ohio University and Michael L. Naraine, Brock University
Abstract
Spectator sport sales staffs have a key problem. Sport sales personnel are important to organization revenue generation, yet the role is highly demanding and highly overtaxing. Employees must balance the exhausting task of pre-qualifying prospects with also executing the duration of the sales process for engaged customers. Long hours, facing daily rejection, and inefficiencies in the pre-qualification phase affect culture, turnover, and the organization’s bottom line. Natural language processing (NLP) in the form of automated conversation with customers is positioned as a solution. This paper describes the capabilities, benefits, and drawbacks of NLP application for creating efficiencies in the pre-qualification phase in spectator sport sales.
Introduction
The Problem
Ticket sales are an important revenue stream for spectator sport organizations (Popp, Simmons, & McEvoy, 2017). Sport sales staffs are tasked with maximizing ticket sales via recruiting new buyers, retaining ticket holders, and upselling current fans.
To be successful, sales personnel must master a range of skills and tasks in the sales process from prospecting to probing needs of prospective buyers to closing the sale (Pierce & Irwin, 2016; Popp et al. 2017; Popp et al., 2019). While sales staff operations and success are important to the organization, sport sales staffs are often highly constrained (Popp et al., 2019).
In the initial stage of the sales process alone, to keep a steady stream of buyers in the pipeline, sales personnel make roughly 75 to 125 calls per day, with most of these calls ending in rejection (Pierce et al., 2017). While facing this rejection, sales employees must balance maintaining and upgrading existing relationships while learning and developing sales skills (Popp et al., 2019).
For instance, sales staff engage potential buyers with phone call follow-ups, text messages, as well as social media messages on platforms like Facebook, Instagram, Twitter, and LinkedIn (Naraine & Parent, 2016), while also developing the skill set for how best to reach individuals on each of these mediums (Popp et al., 2019).
There exists a key problem for sport sales staffs that results in numerous ramifications; these roles are highly important, highly demanding, and highly overtaxing especially for those new to the sport industry (Pierce & Irwin, 2016; Popp et al., 2019).
In this capacity, sales roles can be grueling and particularly difficult for new recruits learning the techniques and nuances required for success on top of the modern demands (e.g., ongoing touchpoints with buyers). As such, sport sales teams are often riddled with high turnover and systemic culture challenges (Popp et al., 2019).
Consequently, the bottom line is affected with inefficient prospecting of potential sales.
Pierce et al., 2017
Sport organizations have realized the extent of this problem (Pierce et al., 2017; Popp et al., 2019). To mitigate the inefficiencies and rejection associated with this initial prospecting stage, sport sales managers employ campaign prioritization (Pierce et al., 2017).
This process is often driven by an analytical ranking system (Newman et al., 2017). Sales managers and analysts separate leads into campaigns, then rank the campaigns according to customer propensity to buy.
For example, one campaign may involve prior buyers, while another may contain prospective buyers with no contact with the organization. Within a given ranking system, the campaign with prior buyers may be ranked higher because of the pre-existing customer affiliation.
The sales staff is then tasked with opening the door with the prioritized leads before moving to the next campaign. But, even within the campaign prioritization process, there exists both rejection from uninterested leads and missed opportunities in the pipeline.
A campaign that is not prioritized may still have interested leads that await engagement (Newman et al., 2017). A campaign that is high priority in terms of engagement will still have leads that are uninterested and will not engage. Given that there are only so many hours in a day, and only a certain number of sales personnel, some campaigns and some potentially interested buyers may never receive a phone call.
Unfortunately, there is also the potential for sales staff to be so unenthused and disoriented from working through the campaign process that buyers who are “on-the-fence”, and could potentially be convinced to complete a purchase are not properly managed and effectively sold.
Even though analytical approaches attempt to mitigate the rejection and inefficiency, there are still leads that would possibly be engaged that are not contacted, and there are portions of the sales staff that will be spending time on low-yield opportunity tasks.
The pervasiveness of this issue has numerous costs for sport organizations that extend from simply missing out on potential development for the business operation to a poor, dreary sales culture. To compound the issue, entry level sales staff may be placed on lower level campaigns to prove their merit toward more engaged leads (Pierce et al., 2017).
For the modern sales staffer who is also using e-mail, text, and phone calls to follow-up with buyers, attempting to increase frequency and communication touchpoints with the unengaged can also have an adverse effect; while these instantaneous technologies bring the staffer closer to the buyer, it can also wane on the former’s psyche (Pierce & Irwin, 2016).
For instance, sales staffers may wish to control their personal social media brand and not want to use it as an aggressive sales tool. However, sales staff frustration can also lead to collective, sales culture implications beyond the individual implications.
An atmosphere of rejection, aggressive pitching to meet campaign quotas, and consistently high turnover (Pierce & Irwin, 2016) creates problems for sport organizations forcing sales managers to take the time and resources to try and alleviate such negativity with retraining whilst recruiting new staff and subjecting them to the same environment (Pierce et al., 2017; Popp et al., 2017). With potential buyers still not being called, there are missed opportunities in the pipeline. This initial engagement effort in the sales process needs reinvention.
Two questions are posed. Is there a mechanism for sales staff to direct their attention toward engaged leads? How do we step past that faulty inefficient, routine pre-qualifying process?
Natural Language Processing: Automated Friend or Foe?
Enter artificial intelligence; well, natural language processing (NLP) more specifically. NLP is one branch of artificial intelligence, encapsulating computer capability to recognize patterns in and reproduce the human language (Collobert et al., 2011; Joseph et al., 2016; Naraine & Wanless, 2020).
While the popular connotation of the word “data” may be a series of numbers in a spreadsheet, the human language qualifies as various aggregations of textual data. Just like statistical tools can recognize patterns in numerical data, algorithms – systematic processes to evaluate data – can find patterns in textual data and be programmed to generate responses in kind.
This ability to engage and respond is critical to the two-way communication mechanism in sport management (Pedersen et al., 2021) and, as Berger et al. (2020) wrote, “words are part of almost every workplace interaction” (p. 1).
The availability of rapid and free machine language translation, computerized social media, and other narrative mining including sentiment retrieval have all accelerated once tedious, time-consuming, and inefficient business processes for a wide range of industries (Hirschberg & Manning, 2015). Regarding customer relationships, businesses can utilize the text that customers produce to predict consumer behaviors, as well as conduct exploratory analyses attempting to understand why customers feel certain ways about a product (Berger et al., 2020).
Sports, events, and festivals are no exception to improvements afforded by NLP; in one example, for the 2014 World Cup, NLP algorithms allowed event marketers to identify the most highly engaged social media posts and to translate responses into six different languages (Ponce, 2014). In another, an NLP-driven concierge hosted and consolidated a once confusing multi-site ticket purchasing process for the Broadway show, Wicked (Satisfi Labs, 2021).
Possible advantages also span monitoring stakeholder activity in the public narrative to computer-generated journalism (Naraine & Wanless, 2020). The overarching gains boil down to creating efficiency, automating routine tasks, reducing human error, and interpreting large volumes of textual data (Deshpande & Kumar, 2018).
Consequently, NLP (in the form of automated conversation) is positioned as a solution for the overwhelming pre-qualification phase of the spectator sport sales process (see Figure 1). The following describes why.
When a salesperson picks up the phone, dials a number, drafts an email, or composes a social media message, they are performing a routine task involving textual data (spoken or written word from a sales representative to a customer and from customer back to the sales representative). However, as previously stated, this textual data can produce volume that becomes overwhelming for sales personnel, as representatives are tasked with opening multiple conversations with 75 to 125 customers daily (Pierce et al., 2017).
Imagine being able to create a system where the first-level communication paradigm did not involve a member of the sport sales staff and instead was supplanted by computer technology. That is exactly what automated conversation is referring to in this context. Chat bots, which are trained algorithms to simulate and automate conversation, can perform the work of opening conversations with an unlimited number of potential clients without being stressed, overwhelmed, or deterred, and without leading to a negative team culture.
Not all chat bots involve NLP. Rules-based chat bots involve pre-programmed responses to a specific set of questions or verbal cues that usually involve pressing a set of buttons or limited choices for proceeding.
One example of how these rules-based bots operate in practice is the standard phone tree which a customer experiences when calling into a customer service hotline. “Press one if you would like to get location hours; press two if you would like to purchase a ticket” are typical variants of the rules-based bot, a system pre-defined by choices without any flexibility or ongoing adaptability.
For some business operations, these types of bots are sufficient at the pre-qualification process by directing the customer to the correct department or associate. It can also alleviate the strain on sales personnel, as only customers who explicitly declare their interest to purchase can be handed-off to sales staff for a more in-depth conversation about customer needs, the product scheme, and price levels.
However, a chat bot that has NLP applied into its process can create a more nuanced, customizable approach. NLP-guided chat bots have evolved into more capable and personalized conversation agents (Hirschberg & Manning, 2015; Fahad & Yahya, 2018). Chat bot algorithms can be trained on a plethora of documented customer conversations to prepare the algorithm how to respond (Naraine & Wanless, 2020).
This programming is a remarkable opportunity because, at its core, the NLP-infused chat bot self-learns over time. Because of NLP capabilities, the chat bot’s algorithm transitions speech to usable data, then translates that usable data back to congruent, seemingly organic conversational responses (Deshpande & Kumar, 2018). The result is a chat bot that can comprehend human communication meanings and can be trained to remember earlier parts as the conversation drifts and evolves with new, nuanced topics and issues.
The advantages over a rules-based chat bot involve a more individualized experience as well, for example, reading a customer’s emotions and responding. The chat bot algorithm can decipher dissent and frustration from the lexical arrangement of words and how they are used. In this capacity, chat bots can also apologize and demonstrate sympathy for consumer concerns.
These types of personalized approaches win in sales (Pierce et al., 2017), and are much more desirable than non-emotive rules-based bots that are confined, restricted and do not make the customer feel welcomed (instead making it feel like they are talking to a robot).
Both Satisfi and Conversica, two prominent automated conversation agent providers, boast success stories serving the sport industry in this manner (Conversica, 2020; Satisfi Labs, 2020). The two companies offer a “virtual assistant” that can communicate through website, mobile text, and social media chat. The bot interacts with consumers looking to purchase tickets to their favorite sports team as well as post-ticket sales including food, beverage and game-day support.
Companies claim a plethora of event hosts as customers including Broadway shows, zoos, teams from the Big 5 North American sports leagues (i.e., NFL, MLB, NBA, and MLS; Conversica, 2020; Satisfi Labs, 2020) and the Melbourne Stars cricket team in the Big Bash League (Satisfi Labs, 2020).
Pushing first-level, prospective conversation to a digital chat bot allows the organization to tap into potentially unreached customers as well as the increased number of fans interacting with brands online (Naraine, 2019; Naraine et al., 2019).
A significant number of conversations can be diverted away from human staff members, freeing up capacity to focus on engaged sales leads and other relationship-building activities.
Incorporating a chat bot to pre-qualify customers seems like a fail proof solution – but not so fast.
While trade literature and company promotional materials covering the topic encourage the applicability of chat bots to resolve inefficiencies in the sales process, peer-reviewed literature covering overviews of NLP also illuminates the downsides (Akbari, 2014; Fahad & Yahya, 2018; Hirschberg & Manning, 2015).
While able to communicate and emote better than rules-based bots, and although there have been substantial improvements in the ways these bots interact (Hirschberg & Manning, 2015), technical limitations still exist (Fahad & Yahya, 2018). The human language is ultimately flawed and complicated. Aspects of normal human interaction can be troublesome for automated bots, such as natural pauses, subtle conversational cues, and cadence with taking turns in conversation (Hirschberg & Manning, 2015).
As sport brands achieve global reach, adding more languages only compounds this problem (Akbari, 2014). Slangs and nuances vary among dialects and change over time. Disruptions in conversation and misunderstandings can reflect poorly on the brand, create mistrust, fail to produce a truly personalized feel, and generate a negative image for potential customers.
While a bot generates the reach human salespeople cannot, the question remains, will the bot reach customers as well as desired?
Sales personnel are trained to embody the sport brand and exhibit charisma (Pierce et al., 2017; Popp et al., 2019). Ensuring a conversation bot represents the brand and displays this natural charm is a challenge (Hirschberg & Manning, 2015). In addition, growing concerns with privacy issues in this digital age test chat bot implementation (Hirschberg & Manning, 2015).
Organizations will have to answer to customers curious about how their conversations may be stored. Furthermore, bots are expensive; professional teams may be capable of absorbing this expense, but the cost may overwhelm other sport industry levels. And while the chat bot can resolve time constraints with the pre-qualification phase, additional time is needed to fully integrate the bot into the organization.
This integration is a new task for sport teams that may not have the expertise nor the understanding of how the chat bot will fit in the overall sport business culture. Less discussed, but ultimately very important, is the set of best practices for management of the bot by the nontechnical personnel working closely with the technology (e.g., the sales team). Orienting employees, monitoring progress, and measuring outcomes all represent critical yet uncharted territory for sport organizations harnessing this new technology. While the upside of incorporating a chat bot into the pre-qualification phase of spectator sport sales is attractive, how a team fulfills the bot’s potential will depend on addressing these challenges.
Conclusion
Sport sales roles are highly important, highly demanding, and highly overtaxing. It makes sense that these positions experience high turnover, culture challenges, and pipeline inefficiencies. Opportunity, however, to reinvent the sales process and mitigate these issues is knocking with the continual development of NLP. Replacing the initial pre-qualification phase with automated and personalized conversation relieves sales personnel of low-yield, frustrating work.
Instead, sales personnel can spend more time with engaged leads in the relationship building phase of the sales process. Bot implementation is not without its challenges; sport businesses face ensuring the bot reflects the brand, regulating privacy issues, reconciling with NLP limitations, and bot coordination within sport organization culture. Ultimately, the advantages and disadvantages warrant the development of best practices in chat bot facilitation specifically for the spectator sport industry.
References
Akbari, A. (2014). An overall perspective of machine translation with its shortcomings. International Journal of Education & Literary Studies, 2(1), 1-10.
Berger, J., Humphreys, A., Ludwig, S., Moe, W. W., Netzer, O. & Schweidel, D. A. (2020). Uniting the tribes: Using text for marketing insight. Journal of Marketing, 84(1), 1-25.
Collobert, R., Weston, J., Bottou, L., Karlen, M., Kavukeuoglu, K., & Kuksa, P. (2011). Natural language processing (almost) from scratch. Journal of Machine Learning Research, 12(1), 2493-2537.
Conversica. (2020). How the Sacramento Kings added millions in pipeline. Conversica.
Deshpande, A. & Kumar, M. (2018). Artificial intelligence for big data. Packt Publishing.
Fahad, S. K. and Yahya, A. E. (2018). Inflectional review of deep learning on natural language processing. International Conference on Smart Computing and Electronic Enterprise, Malaysia.
Hirschberg, J., & Manning, C. D. (2015). Advances in natural language processing. Science, 349(6245), 261-266.
Joseph, S. R., Hlomani, H., Letsholo, K., Kaniwa, F. & Sedimo, K. (2016). Natural language processing: A review. International Journal of Research in Engineering and Applied Sciences, 6(3), 207-216.
Naraine, M. L. (2019). Follower segments within and across the social media networks of major professional sport organizations. Sport Marketing Quarterly, 28, 222-233.
Naraine, M. L., & Parent, M. M. (2016). Illuminating centralized users in the social media ego network of two national sport organizations. Journal of Sport Management, 30, 689-701.
Naraine, M. L., & Wanless, L. (2020). Going all in on AI. Sports Innovation Journal, 1, 49-61.
Naraine, M. L., Wear, H. T., & Whitburn, D. J. (2019). User engagement from within the Twitter community of professional sport organizations. Managing Sport and Leisure, 24, 275-293.
Newman, C. L., Cinelli, M. D., Vorhies, D., & Folse, J. A. (2018). Benefitting a few at the expense of many? Exclusive promotions and their impact on untargeted customers. Journal of the Academy of Marketing Science, 47, 76-96.
Pedersen, P. M., Laucella, P. C., Kian, E. M., & Geurin, A. N. (2021). Strategic sport communication (3rd ed.). Human Kinetics.
Pierce, D. A., & Irwin, R. L. (2016). Competency assessment for entry-level sport ticket sales professionals. Journal of Applied Sport Management, 8, 54–75.
Pierce, D. A., Popp, N., & McEvoy, C. D. (Eds). (2017). Selling in the sport industry. Kendall Hunt.
Ponce, J. (2014). Sony launches social network that will form centerpiece of 2014 World Cup sponsorship. SportTechie.
Popp, N., Simmons, J., & McEvoy, C. D. (2017). Sport ticket sales training: Perceived effectiveness and impact on ticket sales results. Sport Marketing Quarterly, 26, 99-109.
Popp, N., Simmons, J., & McEvoy, C. D. (2019). Effects of employee training on job satisfaction outcomes among sport ticket sellers. International Journal of Sport Management and Marketing, 19(3-4), 147-160.
Rettig, M. (2017). Publisher GiveMeSport uses AI to deliver engaging content to audiences. SportTechie.
Satisfi Labs. (2020). Sports. Satisfi Labs.
Satisfi Labs. (2021). Wicked launches Broadway ticketing. Satisfi Labs.
Sport & Entertainment Review 