Digital Loyalty Programmes: Pull Strategies in B2B Channel Marketing

Digital loyalty programmes are an increasingly common tool for business-to-business marketers hoping to increase repeat sales through deeper customer engagement. In consumer markets, such programmes do little to inﬂuence behavioural loyalty and disproportionately attract the ﬁrm’s existing heavy buyers. Industrial buying, however, relies on direct sales channels and features negotiation and reciprocity. Loyalty e ﬀ ects may therefore di ﬀ er in B2B, and although no clear picture yet exists, such knowledge is important as B2C digital loyalty programmes grow in popularity. Here, the authors describe programme membership’s evolving characteristics over in a B2B scheme that was launched in the US metal-cutting tools manufacturer customer base. Findings are consistent with the idea that the scheme recruited the heaviest buyers earliest and had an insigniﬁcant e ﬀ ect on total revenue. The authors discuss managerial implications, particularly about (1) managing the rollout of similar schemes and (2) refocussing on the programme objectives to maintain sales from the lightest rather than the heaviest buyers.


INTRODUCTION
The chapter considers the effectiveness of end-user digital loyalty programmes in the business to business (B2B) domain.Pull strategies of this type that create demand among end-users are common in consumer markets but unusual in B2B marketing, where push tactics are more normally prioritised to promote products through distributor channels.Advances in data science now make B2B loyalty programmes both affordable and practical, but their benefits may not yet be fully understood in a B2B setting.This chapter aims to set out those benefits by discussing the launch and diffusion of a novel end-user loyalty programme that connected a heavy industrial manufacturer with its end users through the existing distributor network.
The term "loyalty programme" captures a range of marketing initiatives, including discount vouchers, reward cards, tiered service, membership levels, dedicated customer support, and other techniques (Henderson et al., 2011), all designed to encourage repeat buying (Uncles et al., 2003).Loyalty programmes share a common purpose: to attract new consumers to the brand and to lock in existing members (Juetten et al., 2006;Ziliani & leva, 2019), and there is some evidence that they may successfully drive market share growth (Humby et al., 2004).
Previous authors challenge the idea that in consumer packaged goods or retail, schemes targeting consumer loyalty can create differential advantage.This is because it is normal for rival brands to share customers, and those rival brands often also have rival schemes (Verhoef & Langerak, 2003;Liu & Yang, 2009).In addition, the capacity of the heaviest users to increase purchasing is limited (Bolton et al., 2000;Lal & Bell 2003); the schemes appeal mainly to existing, loyal customers (Meyer-Waarden & Benavent, 2006); and loyalty strategies are in any case defensive, doing little to replace lost users (Sharp & Sharp, 1997).In other words, "loyalty programmes are a poor marketing tool" (Shugan, 2005;Magatef & Tomalieh, 2015) which reflect rather than change existing behaviours.
However, important differences exist between buying behaviours in B2B and B2C (business to consumer) contexts.Industrial manufacturers and their distributors have a relatively smaller customer base with bigger and more regular purchasing, some fixed contractual binding, professional buyers and sellers, a close relationship between parties (Jackson & Cooper, 1988;Wilkinson et al., 2016), and an emphasis on customer retention strategies (Dorotic et al., 2012).These differences may affect the operation of a loyalty programme, or render it ineffective.They might also breach the as-if random interpretation of buying behaviour that underpins the generalised stochastic models used to assess and predict repeat buying.
The adoption of end-user digital loyalty programmes in industrial markets would seem to meet many of the objectives of the B2B marketer, encouraging higher loyalty, reduced shipping costs, and increased repertoire size (Capizzi, 2002;Lacey & Morgan, 2009;Kwiatek & Thanasi-Boce, 2019).Here we ask, are loyalty programmes, a pull strategy, an appropriate and effective tool for channel marketers where push strategies are more generally the norm?
In this chapter, we explore three questions.The first -are pull strategies a suitable tool for B2B marketing at all? B2B marketing investment is usually designed to maintain distribution relationships and competitive advantage within the distribution channels.B2B brands are not usually advertised to B2B customers to the same degree as B2C brands because the one-to-many model is not justified by the limited size of the B2B target market and its bulk buying.Instead, relationship marketing is the usual technique designed to reward and develop the behavioural loyalty of distributors and particularly the most valuable ones (Reichheld & Sasser, 1990).But just as in consumer markets, while the end-user buys the brand, they are not directly the brand's customer.
Second, are loyalty programmes useful mechanic in industrial marketing?In B2C and services marketing, some question whether loyalty programmes grow sales.Previous research has extensively applied stochastic models of aggregate repeat-buying such as the NBD or NBD Dirichlet (Dowling & Uncles, 1997;Sharp & Sharp, 1997) to establish that while there may be other benefits, programme membership has little impact on expected purchase frequency or scheme adoption across the customer base (Uncles et al., 2003).Applications of the models and the empirical generalisations they describe have successfully extended to some B2B buying contexts (McCabe et al., 2013;Pickford & Goodhardt, 2000;Schmittlein & Peterson, 1994), but this work has not yet extended to the study of B2B loyalty programmes and in particular, it pre-dates the rapid advance of advanced digital marketing techniques.
Our third question, then, is whether digital advances have changed anything.The greater marketing efficiency noted by Saura et al (2017) has been achieved by the shift to online platforms (Russell, 2010), and this creates two meaningful opportunities.Firms now have access to a vast array of digital tools to communicate with customers, and second, such tools have made marketing activities more measurable by improving marketers' ability to access, gather, analyse, and report data (Pauwels et al., 2009) and take action.But the complex nature of channel relationships means that end-user performance data is often hard to come by, and ownership of detailed market insight is both limited and one-sided.A manufacturer/end-user loyalty programme that operates through the distributor network could address this challenge to the benefit of all players.Therefore, while an online B2B loyalty programme might not drive improvements in sales directly, as a result of the data-driven insights generated, it may improve a manufacturer's ability to manage sales performance through the whole channel.Advantageous distributor relationships might then be developed based on allocations of end-user performance rewards generated and managed digitally and at a low cost.
Following a recent call for further research into the effectiveness of loyalty programmes (Chen et al. 2021), this chapter raised three critical questions.Can a B2B pull strategy based on a loyalty programme (1) solve the problem of customer and end-user management in complex distribution channels (2) build sales through end-user loyalty, and (3) provide strategic customer relationship management (CRM) insights that deliver a competitive advantage to the manufacturer while benefitting the distributor network.
To explore these questions, we offer a case study that describes the introduction and rollout of a novel digital B2B channel loyalty programme in a heavy industrial setting and the unforeseen benefits achieved for all channel members involved.We examined this, first evaluating changes in aggregate behavioural loyalty in membership sales data as the new scheme rolled out and from contemporaneous interviews with management at the global digital marketing agency that innovated the scheme.
In the following sections, we provide an overview of the literature framing the study then a summary of the methodology and data sources employed.The results and their implications are then discussed, with a summary of limitations and suggested further research.

Channel Marketing
Most industrial manufacturers do not directly sell their goods or services to their end consumers.Instead, they turn to market intermediaries (wholesalers, distributors, retailers) to facilitate the flow of manufacturing output to the market (Brocato, 2010), also known as channel marketing.In turn, mar-keting intermediaries work as a mediating link between manufacturers and end-consumers (Kotler & Armstrong, 2010) and can add significant marketing value.For example, intermediaries represent the manufacturer and its brands in the marketplace and may provide logistic support by offering pre-and post-sales services (Fayyaz & Azizinia, 2016).
In turn, intermediaries work with multiple, competing partners (Calantone & Gassenheimer, 1991;Aman, 2017); therefore, a common strategy for rival manufacturers is to push their products through the channel by incentivising their intermediaries through the promotion mix (Armstrong, Kotler & Opresnik, 2011).This approach is referred to as a "push" strategy (see Figure 1), and its primary purpose is to increase repeat business at the expense of channel rivals (Dibb et al., 2012) while favourably developing the channel relationship.
Management may believe that push marketing builds strong channel relationships, which lead to loyalty and mutual profitability (Morgan & Hunt, 1994;Payne et al., 2017).The strategy is clearly beneficial for intermediaries as they receive competing and ever more lucrative deals.However, for manufacturers, while such promotions represent a continuous and necessary investment, they appear to have no persistent positive impact on sales (Nijs et al., 2001).They may even train intermediaries to become deal prone by shifting focus from the brand to the promotions (Scriven et al., 2017).
In B2C marketing, manufacturers who are at arm's length from their users redress this balance by investing directly in consumer-brand relationships.These are designed to create demand -"pull" strategies (Figure 1).Pull marketing differentiates and builds consumer-brand knowledge through advertising, develops consumer loyalty with reward schemes, and attempts to strengthen consumer-brand relationships through continuous online interactions (Lim et al., 2019;Thaichon et al., 2020).Customer loyalty programmes provide a valuable platform for these interactions, and at their core, they share a common purpose: to attract new consumers to the brand, lock in existing members (Dowling & Uncles, 1997;Juetten et al., 2006) and deliver competitive advantage from the data they collect (Grewal at al., 2011;Breugelmans et al., 2015).
Nevertheless, few have ever attempted to include an end-user digital loyalty programme in a B2B setting, largely because of the relatively smaller number of end-users to be served.Significant differences exist between B2B and B2C market (Zinkhan & Cheng, 1992;Liu et al., 2018).For example, impulse buying is unusual because B2B firms purchase raw material to fulfil derived demand.The buying decision depends on professional buyers and sellers (Jackson & Cooper, 1988;Oakley et al., 2021); orders are usually bigger and often more infrequent, sometimes scheduled using specialised software, while the initial sales process often takes considerable time.It may also involve fixed contractual binding and negotiations regarding technical support, credit, and delivery terms (Grewal et al., 2015).There is then an emphasis on customer retention (Dorotic et al., 2012) throughout the channel, and channel effectiveness then depends on leveraging relationships.
An end-user loyalty programme would offer a novel solution to many of the manufacturer challenges with channel marketing in this setting.For example:

Loyalty Programmes
Marketers use loyalty programmes hoping to retain their most valuable existing members.However, previous researchers who examined repeat purchase rates pre and post the introduction of a loyalty programme have found very limited evidence of "excess" behavioural loyalty when measured against stationary market models of buying norms such as the NBD-Dirichlet (Sharp & Sharp, 1997).Like everything else in competitive marketing, they suggest that if one brand introduces a successful intervention, others must follow.In repertoire markets like grocery retail or consumer packaged goods where switching barriers are low, loyalty programmes generally attract those who stand to benefit the most: the heaviest shoppers.Since they already use several brands in the category over the course of a year, they often subscribe to several competing loyalty schemes.If, as Meyer-Waarden & Benavent (2006) conclude, loyalty schemes are simply more attractive to heavier rather than lighter users, then they become a cost of doing business without creating an advantage by segmenting the market, particularly if they do not attract or are not targeted at, lighter buyers.

The NBD Model of Repeat Buying
In established B2C markets, near habitual repeat buying of consumer packaged goods has successfully been described and predicted at the aggregate level (i.e.across the customer base) with zero-order stochastic models.The most commonly adopted of these (Ehrenberg, 1959) is the NBD, which assumes that population buying propensities are distributed gamma across all buyers, and the timing of their purchases is random around a fixed mean, following a Poisson distribution.
The NBD is simple to use to evaluate buying of a single brand or a category when buying is stationary.It is highly generalised and linked to empirical generalisations in repeat buying that has been widely tested and supported in many marketing contexts (Uncles et al., 1995) across North America, Asia, Europe, and Australasia.Recently, the NBD has been extended to different types of behaviour such as gambling (Lam & Mizerski, 2009) and mobile phone service (Lee et al., 2011).While the use of the NBD has been tested across various domains (e.g.Dawes et al., 2020;Trinh et al., 2014;2016;2018), its application in the B2B context has been limited to only a few studies.Some of these have suggested that since B2B customer behaviour is less random (especially within membership contexts), it may constitute a boundary condition (Sharp et al., 2002); thus, the model may not explain B2B behaviour well.
In the study of loyalty programmes, the NBD and its more complex relative the NBD-Dirichlet (Godhardt, Ehrenberg & Chatfield, 1984) provide a suitable benchmark to examine divergence from stationarity.For example, to demonstrate if a particular class of heavier or lighter buyers change in propensity resulting from adopting loyalty programme membership (Sharp & Sharp, 1997) and being rewarded for purchasing more.In any given time period, to calibrate the NBD, only two inputs are required.The penetration of a customer group within its category and the average purchase frequency in the given time.The model output then describes the distribution of buyer classes (1, 2, 3, 4….n), the repeat purchase rate and the customer acquisition rate between periods; in other words, the key measures of behavioural loyalty in the customer base.

Loyalty Programmes in Digital Marketing
In a B2B setting, buyers are not always primary decision-makers.In many cases, professional procurement managers are moved from department to department with the express intent of discouraging buyer-seller relationships from being developed.Therefore, there may be a first-mover advantage for a manufacturer across its distribution channel in owning a B2B loyalty programme targeting end-user relationships to develop sales at the expense of rivals.While any competitive advantage offered by launching the first B2B loyalty programme may diminish over time, the benefit obtained by a constant flow of information about end-customer behaviour would still remain valuable (Palmer et al., 2000), continuing to provide leverage with the distributors.
The role of loyalty programmes has changed radically in recent years, owing to the web's immediacy and data-driven marketing.Today, customer loyalty programmes operate digitally with e-commerce and online payment infrastructure and reward members for various desired actions.The programmes have eliminated the need for physical cards, and digital evolution has made programme adoption much simpler (Smith, 2000).The other benefits brought by new technologies benefit marketers by providing detailed insights (Reyes-Menendez et al., 2018), identifying and targeting customised segments, and providing an interactive platform where customers and brands can see progress towards rewards (Leva & Ziliani, 2016).
Besides, the transition of loyalty programmes into the digital space has addressed some previously stated challenges.In comparison with traditional schemes, Leva & Ziliani (2016) state that online loyalty programmes may have distinct benefits: They can increase single-brand loyalty (O'Malley, 1998), provide a freestanding brand touchpoint (Lemon & Verhoef, 2016), keep the memory of brand fresh (Romaniuk & Sharp, 2016), and intuitively they are good value for money (Leva & Ziliani, 2016).Therefore, online loyalty programmes may be more likely to increase loyalty and achieve the firm's desired objectives.B2B manufacturer's products have a substantial impact on the quality and productivity of their endcustomers' products.In their innovation, they should fully engage with their customers in parallel with channel partners (Wright et al., 2019).In B2B channel marketing, manufacturers have more motivation to create a relationship with intermediaries (Arnould et al., 2003;Sharma et al., 2020) and systematically offer profit sharing in the form of tiered and over-riding discount structures in return for larger purchase volumes (Leung et al., 2020).Compared with B2C, B2B marketing involves large orders from fewer customers, and it is more feasible for the seller to invest in these incentives (Zinkhan, 2001) with the distributor.But in addition, an end-user digital loyalty programme would help the manufacturer overcome high dependence on their distributors for transactional data.By promoting stronger end-user relationships in this way, manufacturers might also regain control over this data and use the insights it contains to manage channel relationships better.
Digital marketing has opened many opportunities for B2B marketers; however, scholarly research in this area is still at the embryonic stage (Vieira et al., 2019).Online schemes capture detailed customer information through various interactions across all channels -also known as big data (Stourm et al., 2020), through the integration of intelligent analytics and communication technology (Saura et al., 2017).Apart from interactive tools to connect with consumers (Chaffey & Smith, 2017), big data helps generate precise insights reflecting customers' behaviour (Pandey & Gudipudi, 2019).The two main limitations to further research here, particularly for stochastic modellers, would both be addressed by digital programmes in B2B marketing: the difficulties in obtaining B2B data from across channel partner sources and the thin volumes of that data because of the smaller customer base sizes (Lilien, 2016).
A principal objective in B2B marketing is to build channel partner engagement through repeat activity driven by end-user orders.The case for a cost-effective digitally-driven pull strategy in B2B seems quite clear if it could improve distributor and end-user acquisition and retention.In this chapter, we describe an investigation into the operation of such a programme over its first-year roll out, to answer the question;

Research Aims and Questions
Existing research has not yet addressed channel marketing issues from a digital perspective (Thaichon et al., 2020), let alone in the B2B context (Liu, 2020).The aim here is to investigate whether knowledge of B2C relationship marketing can help industrial marketers achieve a firm's goals, by exploring how digital technologies can effectively enable B2B pull marketing.Such research would contribute usefully to the knowledge of industrial marketing.
Therefore the following research questions are proposed:

Research Context and Dataset
A dataset provided by an industry-leading tools manufacturer and its marketing agency was used, containing sales data in the customer base over the rollout of an end-user loyalty programme.This scheme is one of the first in the heavy machinery industry and operates through a wide-ranging distributor network across the United States.The programme is live, but the data provided is historic and describes the first year of operation.All corporate buyers are anonymised in the files, identified only by coding, but sales can be associated with programme rewards and viewed in time series.
The programme was initiated with selected customers as a trial in the United States in 2018 then launched to the full customer base in 2019.The available dataset plots the development of member and non-member transactions as the scheme was first offered to a limited number of sellers in a trial, then a bigger pilot, then opened up and rolled out.The complete membership transaction records include the purchases of over a thousand corporate customers across seven product categories sold through a network of 90 distributors for a year.They can then be analysed at a continuous buying level in timeseries aggregated by month or quarter.

Programme Mechanics
The digital B2B loyalty programme offers points to end customers for purchases made in each calendar month that can be redeemed for manufacturers' products or other merchandise such as electronics and shop appliances.Points are awarded based on quantity purchased across multiple categories in a given month.Earnings per piece purchased are reset every month.Typically, customers earn more points for every additional category purchased, though the point's barriers vary across the categories.
The scheme has developed a 4-level customer tier programme with enhanced benefits at each progressive tier, usually accompanied by initial base points, point multipliers and anniversary gift points.This multi-tier scheme clearly distinguishes end-customer status (e.g.aluminium, steel, stainless, titanium) based on past purchase behaviour-two primary reasons are identified for this.First, membership in a particular tier provides customers with a sense of exclusive identity that can be converted into a tangible status attribute (McCall & Voorhees, 2010).Second, the tier system segments customers according to buying class (heavy, medium, light) to provide differentiated rewards (Rigby & Ledingham, 2004).
The reward scheme dashboard measures repeat purchase activity and offer information management and business analytics tools for manufacturers and end-users with an interactive website or application interface (Marjan et al., 2020).The manufacturer monitors business performance by accessing sales trends and real-time redemption information and can aggregate and visualise data from multiple sources.End-users enjoy exclusive benefits and can check their progress towards points and tiers, motivating them to interact with the platform to encourage higher conversion rates and higher revenues.

METHOD AND ANALYSIS
To answer the three research questions, a multi-method approach was pursued to combine numerical measurement and in-depth investigation (Harrison & Reilly, 2011).Johnson and colleagues (2007, p. 123) provide a general definition of mixed-method research: "…the type of research in which a researcher or team of researchers combines elements of qualitative and quantitative research approaches (e.g., use of qualitative and quantitative viewpoints, data collection, analysis, inference techniques) for the broad purposes of breadth and depth of understanding and corroboration".
A mixed-method approach is suitable because analysing a loyalty programme based only on time series quantifiable indicators may not provide the contextual understanding that carries background information and complex assumptions beneath the numbers (Atieno, 2009).The inclusion of a qualitative approach can complement findings with inspiration for new ideas or possible explanations that can help sharpen existing theory (Siggelkow, 2007).
This study utilised a threefold empirical analysis.First, the transaction data for the rollout period was examined to uncover trends and associations within the data.Second, the data was then modelled using the NBD to identify any divergence from expectation, following the method adopted in earlier B2C studies.In the final phase of the study, the quantitative findings were explored in detail with the reward programme agency directors in-depth interviews.
Stage 1 -Initial Observations.To examine the data for any regularities, identify trends in repeat buying, and establish the roll-out dynamics, the sales data was summarised and tabulated in monthly and quarterly time series over twelve months.Run plots were produced to describe the shape of the scheme uptake and trends in top-line sales by value and by order volumes -purchase occasions.A number of loyalty-based marketing metrics were then established for each stage of the rollout: the penetration of programme members in the customer base, the relative average purchase frequencies of members and non-members at each stage, the repeat rates from quarter to quarter and the average product portfolio sizes across the manufacturer product categories.Initial comparisons could then be made simply between member and non-member buying from these observations.
Stage 2 -Model Fitting.In a second research stage, partly to extend knowledge of repeat buying, and partly to establish any unexpected deviations from stationarity, model fittings were made to the data at the final stage of the programme roll out, i.e. at the point that membership was self-selecting by the end-users rather than where it had been offered by the agency in partnership with the distributors to the "best" customers.At this point, the NBD was fitted to each month's customer base sales data to establish its fit, and evaluate if and how repeat-buying at the manufacturer level was diverging from stationarity, perhaps as a result of the scheme.
Stage 3 -In-depth Interview.Finally, in-depth semi-structured interviews were conducted with the client account director at the agency.Interview topics were adapted from the gaps identified in a recent study by Chen et al. (2021), but allowed the discussion to develop in different directions to ensure depth and richness in responses.Discussion focussed on motivations for launching the loyalty scheme and the objectives set by the manufacturer for its performance.The discussion, conducted online, focussed on benefits for channel members obtained through the loyalty dashboard and the data it provides for the client.It focussed particularly on how the rollout of the digital reward programme helped the industrial marketer achieve its relationship loyalty objectives (Ojiaku et al., 2017).
The interview data were analysed under five themes: • Motivations for launching the scheme • Client and agency programme objectives • Benefits to market intermediaries

•
The insight obtained through the loyalty dashboard • Solutions provided by the program to prevailing channel marketing issues.

RESULTS
The analysis proceeded with an investigation of the research questions described in the previous section.

Push and Pull Strategy in B2B Marketing
The first enquiry focussed on the appropriateness of a pull strategy in B2B marketing through a depth interview with the agency account director responsible for rolling out the loyalty programme for the manufacturer.Before moving onto pull strategies, the interview first examined push strategies in general to examine the motivation for the programme.The director suggested that it was still normal, even in the digital age, for: Representatives from manufacturers [….] to have face-to-face interaction with market intermediaries, to make the sales call.
This is still a standard tactic through the channel in a push strategy, practised worldwide, especially with account-based marketing gaining traction in recent years.It is believed to be effective, particularly in upselling and customer retention.However, it is well-known that face to face selling is the most expensive method in the promotional mix due to costs associated with premises, travel, trade offers and training; it requires a large sales force to carry out personal selling successfully at scale, and it is time-consuming and has a limited reach in a given period.
The account director indicated another prominent issue with push strategies, that of data sharing: Typically, distributors have been very cautious of sharing details of their end customers over the fears of manufacturer bypassing the channels intermediaries and dealing directly with the end customer.
He further added: Such practice has not been beneficial for the manufacturer since this incapacitated any honest, direct feedback.
The arguments align with previous literature and highlight concerns that push strategies can be counter-productive for B2B marketers while necessary.Even where other promotional tactics are put in place, these may not translate into continuous repeat selling after the promotion period is over (Scriven et al., 2017).Other specific risks for marketing productivity in push strategies are noted in the literature:

•
Retailers' hearty appetite for trade deals makes promotions expensive to operate with no longlived change in buying patterns (Martín-Herrán et al., 2017).• Large trade-deal expenses reduce manufacturer's profit margins and only permit a limited amount for further marketing activities (Lassar & Kerr, 1996).

•
The focus on securing a one-time purchase hinders building relationship and reciprocity, and the results are likely to be short-term.

•
The shifting market power to channel partners at the expense of the manufactures due to trade promotions; in return, increased dependence on the intermediaries (Aliawadi et al., 1995).• Lack of interaction with end-user and difficulty in obtaining data affect the performance of critical production decisions (Leng & Jiang, 2016).• Limited visibility of channel partner's transactional data decreases the opportunity for category growth.Manufacturers are unlikely to interpret and benchmark behaviour without such information (Wilkinson et al., 2016).
It was then argued that the fundamental challenge from the manufacturer's standpoint is how to compete more effectively and profitably in the B2B channel to achieve the firm's growth objectives, using push tactics alone.
Introducing digital loyalty programmes to the industrial customer base, a push strategy borrowed from consumer packaged goods marketing addresses many of the challenges identified.Loyalty programmes are cheap to set up, operate permanently, unlike price promotions, and accumulate loyalty by building points/ tiers/hierarchies and providing additional information not available before.The account director adds: By offering this new kind of programmes not only do we get increased efficiency of the communication where we can be more frequent, lower cost and more meaningful, we can also give feedback and insights that we could not provide in the past.
Establishing a programme means that the manufacturer firm can engage directly with customers and motivate their behaviours in meaningful ways.Using digital loyalty programmes, they can adopt multi-channel communication strategies (such as online and mobile devices) through membership data to encourage participation and retention.
The interview noted that loyalty programmes in the B2B industry are still a relatively new idea, therefore:

Getting the first-mover advantage enables a firm to build a relationship quicker and deeper understand what customers want before competitors enter the arena.
Having a unique platform that rewards buying gives the firm a competitive advantage.In summary, pull strategies appear to offer a route to sustainable competitive advantage -as the following example shows:

Primarily because technology and innovation have opened numerous opportunities for operating the platform for the industrial clients.
The benefits are threefold: "…the manufacturer benefits from consistent customer interaction and new insights into end-user behaviour.Distributors can incentivise category expansion and retention, allowing improved marketing strategy and the end-customer relish exclusive benefits and recognition from both manufacturer and distributor".

Effectiveness of Digital B2B Loyalty Programme
To answer the second question is a loyalty programme an effective pull tactic in B2B, the available dataset was explored to evaluate the buyer behaviour outcomes as the programme was rolled out.Sales made to scheme members were analysed in three quarters over the first year of the programme operation, corresponding to the launch to widening customer base segments in three phases.Monthly sales and scheme membership remained largely stationary within each phase, and Table 1 describes the main buying metrics observed.In the left-hand section, the volume and value analysis of orders is shown, and on the right, the aggregated behavioural metrics; buyers in each period, and two loyalty metrics, average purchase frequency and month to month repeat rate of buyers.
The table shows that programme sales grew rapidly as the scheme rolled out; from just $21,000 in its first quarter, to over $800,000 in its third quarter.The table also shows that not only did the scheme sales value grow, the average value of individual invoices to scheme members also increased by over two times.
Clearly though, looking at the right-hand side of the table, membership is the driver of programme revenue growth.The data show a dramatic and rapid fall in average purchase frequency, while sales and buyer numbers are highly correlated.Moreover, as the membership base grows in size, it holds a steady monthly repeat purchase rate, such that on average, around 93% of buyers in the membership in one month repeat in the next.
The membership is drawn from the existing buyers of the brand, and from across distributors.In the third window analysed, there are 166 buyers, which is 16% of the customer base of end-users.The findings are entirely consistent with the idea that the scheme is recruiting the brand's heaviest existing buyers.A near 100% repeat rate, a doubling of the invoice value, and near-stationary buying characteristics in each window all suggest little room for improvement in behavioural loyalty!The decline in purchase frequency is curious but may indicate either (1) a degree of order-bundling to obtain scheme rewards or (2) simply that the very heaviest buyers were recruited first.Further analysis was then undertaken to investigate the scheme in the context of the firm's total customer base.
First, the investigation was extended outside the scheme membership data to examine the shape of total brand sales through all distributors.Figure 2 shows a monthly sales analysis in time series for the complete brand customer base -non-members and loyalty programme members -to identify any performance trend between July and January 2000.Across the period, total sales appear to be broadly stable -excluding a seasonal Christmas dip -meaning that although scheme sales are increasing, they are taking a bigger share of a stable whole.On the other hand, it is also possible to say that overall manufacturer sales are maintaining their value in the face of competition.
Whether or not the loyalty programme has built additional sales for the brand has been examined in light of behavioural data; the results are mixed.One apparent reason is that sales remain stable in the customer base data; it is still hard to untangle the scheme's performance from the top line sales curve alone.This is because, under the surface of stability, significant buying fluctuations for different buyer classes and categories may exist that are worth exploring.An analysis of the total customer base would identify significant departures from the norms of stationary buying and, in comparison with the loyalty metrics of scheme members, whether the scheme was influencing loyalty in unusual and unexpected ways.
For example, there is some evidence that the scheme was changing buying patterns by encouraging order bundling -bigger but less frequent orders.The qualitative research had noted that: …customers were bulk buying to get better rewards at a reduced cost on freight.
This might explain the declining purchase frequency because of less frequent orders and reflect increased efficiency due to the programme.
But fitting an NBD would establish whether the loyalty (the distribution of heavy and light buying across the buyer base) was anything other than what might be expected in a stationary situation.Given the stability in the top line sales data, NBD fitting was deemed appropriate for each month, and the estimations are summarised as an average monthly result for the final quarter of the analysis in Figure 3.The model output (shown in the histogram) is a surprisingly close fit to the observed data (shown as the continuous curve).This, therefore, reflects several important characteristics about repeat buying in the B2B customer base and implies that the brand buyers are behaving as expected from the evidence presented in many studies.Notably though, the behaviour of the customer base as a whole is somewhat different from that of the programme members.
For example, where the membership buys every month (a 94% repeat rate on average), only 40% of the total customer base buy in any given month.The repeat rate is also far lower within each month, where the average customer purchase frequency in the membership is 14 orders each month; for the total customer base, that average rate is only 9.9.
The distribution is typically reverse-J shaped so that most buyers are lighter buyers than average: the most common order frequency is only once a month, the buying rate for nearly a quarter of the monthly customers.Therefore, the main finding is that programme members are not typical customers.In addition, the total customer base has not been skewed by particularly heavy buying to vary from the expectation.In short, there is evidence that the scheme may have increased order values for certain buyers, but it has not changed the distribution of purchase frequencies beyond what is expected.
Therefore, while a digital loyalty programme may be an appropriate pull tactic in B2B, it may not necessarily be an effective tactic if the objective is to adjust the loyalty curve's shape permanently.On the other hand, if the B2B customer base is similar to the B2C customer base in its repeat buying behaviour, the programme may deliver some unexpected benefits, which we discuss in the final section of the chapter.

Advances through Digital Marketing
In answer to the third question, we asked whether advances in digital marketing could now make such programmes more widely applicable.These advances have created two critical opportunities for B2B organisations.First, a wide array of digital tools are available to target different customer segments.Second, those digital tools in combination with data science have made marketing initiatives more measurable by allowing marketers to access, gather, analyse, and report data (Järvinen et al., 2012;Pauwels et al., 2009;Saura, 2020), shaping knowledge-based decisions.
Since their inception, loyalty programmes provided relatively inexpensive means for firms to collect information about spending patterns (Hamilton & Howcroft, 1995).However, many traditional loyalty marketers did not seek to obtain transaction data in the past but simply rewarded behaviour in proportion to customer purchasing.Such programmes are less expensive to manage (because of no marketing dollars spent on the analysis) and hinder the opportunities to gain valuable customer insights, develop relationships, and have been seen as more tactical than strategic marketing tools (Palmer et al., 2000).
In order to inform our research question, we asked the agency whether new digital B2B loyalty programmes are more strategic as opposed to tactical.And if so, in what ways?The response: "They are both -we can lay a strategic framework for long-term relationship building but can also use tactical promotions and offers to drive behaviour".With the digital advances, loyalty programmes have allowed the B2B firm more customer base segmentation opportunities because of the data it provides.The detailed analysis of customer profiles helps to identify different buyer classes and their distinct buying behaviour.Moreover, the manufacturer can now use this information to adapt and customise their product offer for individual segments."…to increase category usage and consumption across the customer base, we set customised points and bundles together to engage and upsell to the segments had not actively bought before".
That is, the nature of the data delivered through loyalty can be used to improve performance and develop distributor and customer relationships.
In addition, digital loyalty programmes work as a useful tool for testing the effectiveness of different promotions much quicker -

"It is easier to test and identify what offers customers respond to best".
The firm is now also able to have multiple touchpoint interaction.An example from account director follows: "We can now use the desktop, laptop, mobile, all social media channels to engage with the customer".The marketers can spot the need for quick insight, "...if certain SKUs seem to be popular in one segment and not the other, instant data analysis can answer why the behaviour is happening the certain way".
The analysis can be cross-referenced over time and extend comparison by customer segments, size, geography and industry.The key to effective strategies will be learning from the loyalty programmes analytics platform (Saura et al., 2017) and understanding how to effectively target communication and manage channel offerings to maintain repeat purchases.
But the analysis of the customer base made here demonstrates the critical limitation of the scheme.The data analysed is only pertaining to the behaviour of the scheme membership and not to the nonmembers.While it may seem intuitively essential to manage the repeat buying of the most loyal custom-ers, it may well be the case that they may find it hard to order any more often!Total sales also depend on the contribution of the lightest buyers, those least likely to be in the scheme.
The data analysis highlights the importance of a new focus, not on the repeat buying of the heaviest, but the acquisition of the lightest buyers into the scheme.Regardless of operating in a B2C or B2B market, brand growth usually comes by attracting many light buyers to the brand (Warc, 2021), and brand decline happens due to losing many light buyers.In this regard, the digital scheme could provide an excellent mechanism to interact with light buyers and incentivise their purchases, to reduce churn.If appropriately managed, the B2B marketers' vision should maintain heavy buyers but encourage light buyers to enter and stay in the scheme as a low cost means to "stay in touch" between purchases.This is more easily managed in B2B than in B2C.
Digital innovation provides additional opportunities to improve organisational learning, and leverage end-user knowledge (Abrell et al., 2016).For example, the online platform gathers information that can be used to develop a database to recognise buyer behaviour patterns (Zinkhan, 2001) and predict future purchase rates by incorporating the NBD.Most online platforms have integrated web analytics systems that analyse and report data to help understand users' activity (Saura et al., 2017).By studying corporate buying in this way, a loyalty platform can recommend future purchases and encourage upselling in other categories with customised offers and rewards.
The transition into the online platform and dataset availability has brought several benefits.Digital marketing is a catalyst for channel management (Dasser, 2019) because it builds new engagement networks that stretch back from end-user to manufacturer and allow a deeper understanding of the relationship potential.For example, the loyalty programme seems to be a cost-effective way of maintaining customers.Also, digital has opened lines of communication between a manufacturer and their end-users by providing means to communicate directly to members (Zhang et al., 2019) via applications, emails, direct messages, social media, push notifications.This has eliminated the reliance on intermediaries for customer information and transaction assistance that was not possible using traditional push strategies of the past (Mudambi & Aggarwal, 2003).

DISCUSSION
This chapter analysed the launch of a novel B2B channel strategy supported by advances in digital marketing -we set out to find answers to three research questions: 1) Is a pull strategy appropriate in B2B marketing?2) Is a loyalty programme an effective pull tactic in B2B? 3) Is this now possible by applying digital marketing practices?The research used sales data from a newly launched loyalty programme and in-depth interview with agency directors to answer the questions.
For many, the appropriate marketing strategy in B2B sales involves push tactics to move goods through distribution channels towards end-users by squeezing competitors out of the way.Typically, these tactics include short term volume and price promotions, face to face interactions, and sharing of resources in order to build relatively stronger distributor relationships.These tactics are likely to produce volatile and short-run outcomes because they are usually matched by the rival brands in the same channel, who then regain lost ground.In B2C markets, these efforts are augmented with pull investments designed to create end-user demand, usually advertising or promotions, and again often with only short-term results that serve only to maintain a long-run equilibrium rather than persistent growth.Pull investments are ap-propriate in B2C markets where efficiency is gained through one-to-many investments in the context of "arm's length" customers.Our first question is whether the same tactics are appropriate in B2B markets.
An advantage of the loyalty programme as an end-user pull strategy is that it is designed to build a competitive advantage for the brand in the form of incentives, and those incentives are related to performance.That means that marketing investment is more closely aligned with profitability and, therefore, more efficient.A second advantage is that the marketing pressure applied is continuous in the context of a complex marketing channel where a manufacturer has become separated from the end-user through a distributor network; a loyalty programme makes excellent sense if it rebuilds links with end-users and provides a means to incentivise the intermediaries in addition.It resolves precisely the same challenge faced by B2C marketers, as our interview has demonstrated.
Second, we used a model associated with B2C marketing to evaluate the behaviour of the B2B customer base.We showed that the early adopters of the programme were the heaviest buyers in the customer base, and this is hardly surprising given they have the most to gain.Nevertheless, membership is relatively low a year after launch, and across the total customer base, the scheme appeared to have left the regular distribution of purchase frequencies unchanged.Although the loyalty of heavy buyers might seem to be immune to management, the fact that the NBD gave such a close fit points up a crucial lesson: every customer base has far more light buyers than heavy, and in total, they contribute an important sales volume.
Sales from light buyers are neither optional nor dispensable.In many ways, this presents a more complex marketing challenge than that posed by the few heaviest customers.Brands grow and decline by gaining and losing large numbers of the lightest buyers, and yet here; the loyalty programme is not reaching this part of the customer base.If it did, it would offer a cost-effective solution to the problem of keeping the brand mentally available in the mind of the industrial buyer, particularly where the purchase interval is prolonged.It should help the brand stay in touch with all of its customers.A loyalty programme is uniquely placed to implement this strategic pull task because it is a continuous mechanism and not a campaign-based tactic.
Finally, we asked if the advances in digital science would make the implementation of a reward scheme more efficient.Whether in B2B or B2C, managers look for new ways to increase business efficiency (Berry & Rondinelli, 1998;Abubakar et al., 2019).Interventions through data-driven loyalty programmes allow opportunities to analyse customer base data and take prompt actions, test promotional effectiveness, segment buyer classes.These schemes provide means to interact not only with heavy but also the infrequent buyers.Moreover, an online loyalty programme creates an owned, incremental and engaging brand touchpoint relevant for all customers (Lemon & Verhoef, 2016) even if it is only actively targeted at the most loyal.Engagement with scheme by heavy and light buyers is refreshing brand memory structures (Romaniuk et al., 2013) even if a purchase does not immediately follow, maintaining brand salience among the membership.Online loyalty programmes should be broadly not narrowly targeted among the customer base to achieve higher reach for the brands they promote.This would more likely contribute to the overall brand growth (Sharp et al., 2009;Sharp & Sharp, 1997).

THEORETICAL IMPLICATIONS
The NBD fit to the B2B loyalty programme context suggests that the model is robust and practically useful here.In addition, its reverse-J shape reflects the fact that most buyers in B2B as in B2C are far lighter than the average purchase rate.Consistent with previous loyalty programme literature, the NBD fit proposes that industrial purchasing follows similar repeat buying patterns as B2C, contrary to some prior commentary, which has suggested that regular timings of orders might constitute a theoretical boundary condition (Sharp et al., 2002) This chapter extends the application of the NBD model, which has seldom been applied in manufacturing and industrial goods markets before (although one recent exception is Wilkinson et al., 2016).The findings extend knowledge of several important buying characteristics to the industrial customer base and imply that the brand buyers behave as expected from many similarly close fittings.Particularly here, though, the behaviour of the customer base as a whole is to a certain degree different from that of the membership base, which in this case appears to consist only of the brands heaviest buyers.
We also highlight an increase in business efficiency as B2B customers began to bundle their purchases into bigger but less frequent orders to save shipping costs in order to enjoy exclusive rewards.Thus, the scheme motivated some end-customer behaviour; however, overall buying propensities remained essentially unchanged.
By capturing interview insights, we acknowledge managerial objectives and motivation to launch the scheme to the industrial customer base, its benefits to the channel partners, and the solution pull strategies can provide to industrial manufacturers.The chapter value contributes to understanding the use of big data insights obtained through the loyalty dashboards that support manufacturers to innovate and overcome prevailing channel marketing issues.Moreover, such innovation can help the B2B firms to identify commercially valuable patterns to supply critical knowledge of channel partners and transform customer experience at the end of the value chain -i.e. the end-customer (Wright et al., 2019).

MANAGERIAL IMPLICATIONS
B2B firms' nature towards marketing planning and decision making based on intuition, flexibility, and experience may be challenged by the formalised nature of digital loyalty programme data and the analytics it combines (Armario et al., 2008).Going forward, managers need analysis and interpretive expertise to create unique insights into consumer behaviour to reap the maximum benefits of programme launch.
The key insight here is to recognise the value of light buyers and adapt the scheme to offer incentives to this group, which make the scheme relevant.In line with current thinking, digital loyalty programmes are a flexible, relatively cheap and efficient marketing instrument.
The digital B2B programme allowed the manufacturer to reach the distributor's customer and worked backwards to the end-user.This generated data-driven insights for the manufacturer, offered rewards for the end-customer, and created a demand pull.In conclusion, B2B managers should not hesitate to launch loyalty programmes as they make channel marketing more efficient and facilitate technology-enabled dialogue between the manufacturer and the end customer.But they should be managed for the whole customer base, not the few heavy buyers that constitute around 15% of it, enabling attractive rewards and offers targeted towards the lightest, most infrequent buyers.
In summary, the initial evidence has highlighted the value of digital loyalty programmes for B2B marketers.Managers must consider loyalty programmes performance in the round: do competing manufacturers have rival schemes, can they provide more opportunities for upselling and engagement among channel partners, do they have the potential to manipulate order size and reduce costs through bundling?
The actual value for B2B marketers is not in creating data warehouses through digital schemes from which they can mine potentially helpful information (Palmer et al., 2000).However, it lies in using big data in analysing and identifying commercially valuable patterns or relationships to supply critical competitive knowledge to all channel partners.Knowledge of these recurring patterns helps management because, in the form of empirical generalisations or simple models, they can predict the likely outcome from a given level of investment.

CONCLUSION
Loyalty programmes may be seen as old fashioned, and they have been frequently criticised in the literature as doing little to manage loyalty which generally follows predictable patterns.This study examined the rollout of a new and innovative digital loyalty programme in an unfamiliar context, a heavy industrial setting where a major manufacturer reaches the end-user through a distributor network.
We found that a familiar model of consumer behaviour, the NBD, worked well to describe the distribution of purchase heterogeneity in the B2B brand's customer base.However, its assumption of stationarity remained intact -in other words, the scheme was unlikely to have changed the buying propensities of existing customers.
On the other hand, this drew attention to another aspect of industrial buying, the predictable nature and importance of the lightest brand users.A digital loyalty programme makes an excellent and appropriate tool to deal with the vital marketing task of maintaining pressure on this large group of customers at a low cost if it can be diffused across the whole customer base.

LIMITATIONS AND FUTURE RESEARCH DIRECTIONS
This study produces novel insights into the value of digital loyalty programmes in the industrial buying context.A number of limitations suggest that further research is required that builds upon the results and addresses gaps of the study.
The first limitation stems from the data used-the analysis comprises smaller and inconsistent B2B datasets between programme membership and the entire customer base.Future researchers could incorporate rich longitudinal digital data to help assess schemes' ability to build sales and loyalty for an extended period.Assessments can also be extended to loyalty website and mobile application interface to test whether engagement is predictable (Graham et al., 2021) and if the expected engagement can be modelled on long-term repeat activity to compare against the NBD norms.Such research will contribute to online behaviour knowledge, specifically regarding the predictability of acquisition and engagement to online loyalty programmes.
Secondly, the conclusions are limited to a single brand loyalty programme.In order to test scientific validity and strengthen the generalisability and robustness of the NBD model, future research is now desirable to conduct a replication study at another, or better, at several industrial loyalty schemes simultaneously.Longitudinal B2B membership data reporting usage occasions from a wide range of firms would be preferable.If available across the B2B industry, this data would allow a greater number of brands to be studied; and a category-level NBD-Dirichlet to be fitted, from which benchmarks for a wider range of behavioural metrics could be estimated.Such analysis is crucial for marketers better to understand the nature of competition in their market.The NBD fits well in extended time periods of continuous buying (Dawes et al., 2020), and this warrants extension from B2C into B2B.
Lastly, further work could consider the privacy concerns associated with loyalty programmes data that have received little academic attention to date.Digital loyalty programmes have opened doors of incredible opportunities for B2B managers with the rapid growth of big data, market intelligence and the precision of micro-targeting they provide (Donnelly et al., 2015).By nature, such programmes gain insights from the large amount of data they collect daily; however, not much is known regarding endcustomers willing to share their activity information to receive personalised offerings and communication (Saura et al., 2021).Gaining such understanding is crucial as firms increasingly pay attention to satisfying member's privacy rights and regulators strictly scrutinising ways marketers approach targeted communication in some industries.A recent example is regulations imposed on bookmaker firms in Great Britain to control the use of the VIP schemes data (Gambling Commission, 2020).

Negative Binomial Distribution (NBD):
The NBD theory is a mathematical model used to predict future purchase patterns for a single brand or category from data on purchase frequency and penetration for any given period (Ehrenberg, 1988).
Penetration: The proportion of all buyers under investigation who buy brand x at least once in a period, measured in percentage (Wilkinson et al., 2016).
Pull Strategy: A promotional strategy focuses on facilitating communication flow directly from a manufacturer to end customers to create a demand "pull" for products suitable for their needs (Brocato, 2010).
Purchase Frequency: The average number of purchasing made by those who purchase at least once in a period (Wilkinson et al., 2016).
Push Strategy: A traditional trade marketing strategy used to "push" manufacturers products through the channel by incentivising market intermediaries rather than end customers through the promotion mix (Armstrong, Kotler & Opresnik, 2011).
Relationship Marketing: Marketing efforts designed to establish, retain and enhance long-term relationships with a firm's existing customers to increase profitability (Grönroos, 1994;Peterson, 1995).

Figure 1 .
Figure 1.Example of the pull strategy versus the push strategy in the B2B market Adapted from Sandybayev (2019).

Figure 2 .
Figure 2. Evolution of sales revenues in the total customer base* Customer base * (n) = 1042 *refers to both members and non-members.

Table 1 .
Summary time series analysis of sales and loyalty metrics