Boosting productivity in Canadian banks: a focus on SMEs

Lever 1.1: Optimize how clients are segmented and transitioned between small business and commercial portfolios

It is critical to start off the customer relationship with appropriate segmentation. This has productivity impacts: it right sizes journeys, optimizes cost-to-serve and allocates coverage to high-revenue opportunities.

Many institutions struggle to effectively segment their small business and commercial customers. Segmentation decisions are typically made based on narrow data variables, revenue and loan size, which oversimplifies client needs and can lead to unsuitable segmentation.

Other challenges include the absence of effective referral rewards, leaving relationship managers to take on deals outside their segment to meet sales objectives, and a desire to avoid regulatory compliance activities such as a personal guarantor requirement in small business. This results in some customers engaging in overly complex and highly manual processes, while others face an experience where there is insufficient engagement.

Institutions looking to operationalize a segmentation strategy that boosts productivity should consider the following leading practices:

• Enrich the segmentation framework beyond loan size: Leading institutions are deepening their segmentation frameworks with variables that predict customer needs and service requirements beyond loan size. This includes variables such as industry, lifecycle stage (e.g., startup, growth, maturity) and behaviour (product usage, delinquency).
  
  For example, although a small business with significant international payments and receivables may be smaller based on revenue and exposure, they tend to have complex payment and cash management needs. An institution with mature segmentation capabilities may place this client in the commercial portfolio.

• Enable the front line to identify preliminary segmentation: When a client walks into a branch or accesses banking services through a digital channel, there should be a process in place to identify whether the client should be served at the branch or referred to a commercial representative by asking a handful of simple questions. Institutions can enable staff to perform this preliminary segmentation by providing a playbook on the questions to ask and actions to take.

Lever 1.2: Adopt emerging service coverage models based on customer needs, revenue-generating potential and cost to serve

Once the customer has been properly segmented, it is important to serve them with the appropriate service model.

There have traditionally been two contrasting coverage models: on one end, there is a relationship-managed (in-person) model, often ideal for clients that require highly personalized service through a dedicated relationship manager they can meet in person, where the opportunity to grow share of wallet more than offsets the higher cost to serve. On the other end there is the digital self-serve model, ideal for clients who prefer a digital client-driven journey and the use of in-branch support as required, which allows for a reduction in cost to serve.

As illustrated below, financial institutions have the opportunity to rethink this coverage model dichotomy and consider how they can operate on a continuum of coverage model options.

Between these two ends of the continuum, financial institutions can consider coverage models such as:

• Virtual specialist model, which is an inbound coverage model for small business clients without an assigned relationship manager. The specialists are equivalent to a specialized contact centre resource that can be reached by a branch advisor patching the advisor into a client conversation or with the client directly making an inquiry via online banking. These specialists can cover an array of topics, such as product questions and onboarding support for collecting and submitting the required customer documentation.
• Virtual relationship manager model, which is similar to the in-person model, where the RM manages a portfolio of small business clients, but they’re not bound by the same geographic constraints. This is ideal where there is a fit for a managed relationship coverage model and the client prefers digital channels.

Lever 2.1: Increase auto-adjudication throughput

As institutions seek to make processes faster and smoother, the adjudication process presents a significant opportunity to boost productivity.

In our experience, we often see customer portfolios resemble the “80-20 rule,” where the majority of revenue is driven by a fraction of the customer base. In turn, these portfolios have a “long tail” characterized by a high volume of smaller customers with relatively simple applications. 

We tend to see larger institutions able to auto-adjudicate greater portions of this long tail, enabling their front-office and credit risk staff to allocate a greater portion of their time towards more complex credits. Leading institutions are pushing the capabilities of their auto-adjudication systems, allowing for decisioning of higher complexity applications and loans of up to $1m to $1.5m.

Lending institutions across the Canadian landscape vary in their auto-adjudication maturity. However, there are several common items that institutions of varying maturity are prioritizing:

• A refreshed portfolio risk appetite strategy, including appropriate default definitions and targets, while aligning policies across levels of SME banking.
• Enabling fit-for-purpose technology, including a configurable decisioning engine, expanded digital self-service capabilities and managed workflow through a loan originating system. A configurable decisioning engine prevents decisioning rules being hard-coded into workflows and not maintained, which results in greater volumes of simple, low-risk applications being triggered for manual review. Expanded digital self-service capabilities allow for consumers to experience seamless interactions with their financial institution, on their schedule, providing greater autonomy when managing their banking needs. A loan originating system that enables a managed workflow increases transparency of where the lending application is within the process, and will notify parties to action items. When implemented, duplicative data entries are reduced and users are able to quickly identify the status of the loan. 
• Enhancing credit decisioning models and data, given that internal SME data used for adjudication analytics often lacks the volume of retail portfolios and the resources to maintain ideal data quality. Institutions have an opportunity to enhance their decisioning data through centralized datamarts that include richer internal data (e.g., transactional data) and external data (e.g., multiple credit bureaus). Leading institutions supplement their risk rating and decisioning with alternative data specifically suited for SME lending. They also take into consideration both probability of default and profitability in the credit decision, as opposed to just probability of default. 
• Aligning organization and people enables the right behaviours and new ways of working, such as establishing greater role consistency, clarifying role responsibilities, and developing a change management function to support these transformation programs.

Lever 2.2: Optimize the application and underwriting process

Institutions can optimize the application process by considering what, when and how they collect and process information:

• Consideration 1: What information is being collected?
• We are seeing institutions reconsider whether the information they require their customers and staff to collect and whether the value of this information for risk decisioning is justified. The common result from this exercise is a simplified application that requires less staff time to conduct and improves operational efficiency.
• Consideration 2: When is this information collected?
• Optimizing when information is collected can increase straight-through-processing and minimize the effort that staff spend on “red flag” applications. For example, leading institutions are keeping simple screening questions up front to avoid their staff spending time on applications that cannot be approved under any circumstance. Other information that is necessary, though it requires manual processing, such as collateral verification and select KYC tasks, is being sequenced as pre-funding conditions where possible. Backloading these requirements as pre-funding conditions enables institutions to provide their clients with faster credit decisions.
• Consideration 3: How is this information being collected and processed?
• Institutions are automating various tasks throughout this process where possible, such as KYC requirements, automated spreading tools, and integration between their CRM and loan origination systems to pre-populate applicant information. Leading institutions are also exploring opportunities to centralize their underwriting, which allows them to reduce costs, improve cycle times and enable smaller teams and districts to access industry-specialized underwriting resources.

Lever 2.3 – Employing Artificial Intelligence to create efficiencies in front- and back-office tasks

In recent times, artificial intelligence (AI) has become a relevant topic in the financial services industry. A 2023 EY survey reported that 99% of financial services leaders say their organizations use AI in some capacity. 

While AI has been used in the industry for some time, the use of  generative AI (gen AI) has increased in recent years, with both large and small institutions investing in the technology. Of 151 banks surveyed on the use of gen AI in commercial, small business and retail banking, 60% of large banks have already invested in the technology and 86% of small banks are already investing or planning to invest. Inefficient manual processes are top of mind when banks are investing in this technology: 78% of banks are primarily driven to invest in gen AI to enhance productivity. Additionally, improving customer experience and reducing costs are significant factors, motivating 60% and 59% of banks, respectively, to invest in this technology.

Currently, AI can provide quick responses to a variety of tasks without understanding the context of the task or the situation. While the technology will become more robust in the future, there are a variety of current use cases that employ the technology to create an impact across all value chain segments: 

- Credit scoring: Implementation of AI allows for more variables and data points to be incorporated into the scoring decision. This allows credit risk to have a better understanding of the applicant, customer, industry historical trends and likelihood of repayment, which in turn enables financial institutions to make better lending decisions.

- Operational automation: Operational tasks that are currently manually completed can be automated through gen AI to save time and reduce errors. Such opportunities include back-office functions such as document preparation and loan disbursement, as well as front-office support functions such as KYC, client onboarding and policy/procedure answers

- Contact centre: When a client calls a financial institution’s contact centre, gen AI has the ability to have an effect during all parts of the call. It will first understand the call to select which agent is the best to speak with the client, offer suggestions to the agent in real time and complete a summary of the call. Gen AI will not replace the agent involved, but rather augment the client experience.

- Relationship managers: Gen AI has the ability to assist relationship managers in all parts of the client pursuit and credit-writing process. Before meeting with a client, gen AI’s ability to complete industry research and prepare material for the meeting allows the relationship manager to be more efficient with their time. During the meeting it can provide real-time suggestions to the relationship manager and in a follow-up can assist in completing the lending application. Using gen AI to augment the status allows the client to have a better experience, while also enabling the relationship manager to use their time efficiently.

Currently, financial institutions are predominantly purchasing gen AI solutions directly from vendors as they do not currently have deep data science capabilities in house. Financial institutions take advantage of gen AI through software and the use cases are defined by the software functionality. Over the long term, financial institutions may seek to hire resources with the skills to develop, implement and maintain the technology, which would allow for the deployment as needed rather than relying on out-of-the-box providers.

Lever 3.1 - Optimize pricing with enhanced SME data for analytics and performance measurement

Pricing is a fundamental lever to driving profitable growth, and yet institutions are not consistently pricing optimally and leaving revenue on the table in the SME segment. Financial institutions generally follow an approach where pricing differs as the exposure limit, deposit amount or customer gets larger and more complex.

At one end of the continuum is the retail portfolio, where pricing is primarily centred on product, and at the other is the high end of commercial or corporate banking, where the pricing is highly customized to the customer.

Small business and the small end of commercial banking sit in between and generally use a pricing approach that is predominantly based on product, with some tailoring based on customer attributes. There is an opportunity to optimize data-driven pricing approaches in two key areas in small commercial and small business.

Opportunity 1: Delivering the required data and analytics capabilities

Delivering an effective pricing strategy first requires a strong foundation of data and modelling capabilities. This is a challenge, particularly for smaller institutions, which have lower deposit and lending volumes, which has implications for downstream analytics. As well, smaller institutions struggle to attract and retain analytics talent, with limited investment funding and competing priorities.

By contrast, while larger institutions have access to higher data volumes, they often struggle to maintain data cleanliness in the small business segment, as the retail segment is often prioritized due to its size and relative profitability.

Another important part of an effective pricing strategy is the use of third-party data sources. For a long time, financial institutions have used traditional third-party data sources such as credit bureaus. However, institutions are starting to wake up to the opportunity of employing alternative third-party data sources, particularly as FinTech data aggregators are gaining traction in the Canadian market. The incorporation of alternative data in credit assessment has the potential to increase the efficacy of default predictions, enabling more applications to be approved with more attractive pricing.

Opportunity 2: Enabling front office operations

Another key element of a pricing strategy is enabling the front-office operations. Common challenges institutions face include: 

Discretionary pricing authorities, where relationship managers frequently provide discounts fearing all customers are price sensitive. 

Fees leakage, where fees are not charged. 

A limited view of customer-level profitability, meaning that relationship managers are not taking the whole customer relationship into account when making the decision to provide a discount. 

Limited systematic controls to establish guardrails and nudge relationship manager behaviour. 

Institutions can enable front-office operations through:

• Enhanced controls: Institutions across the Canadian landscape are modernizing their origination systems, and thus have an ideal opportunity to implement a robust pricing control framework. This includes preventative controls that limit discount authority, as well as detective controls that monitor for leakage at the employee and portfolio levels. As institutions mature their data-driven pricing capabilities and monitor the impact of their pricing in the market, they can adjust their models and discretionary authorities. 
• Profitability-aligned performance measurement: Supplementing traditional performance incentives — such as new customer acquisition or new loan exposure — with metrics targeting overall customer profitability, and revenue measures can align relationship manager incentives towards offering optimal pricing.
• Price sensitivity insights: An analytics tool that provides insights on predicted price sensitivity can help relationship managers avoid overestimating their customers’ price sensitivity and offering unnecessary discounts. Price sensitivity analytics can also be used during product development (e.g., conjoint analysis) to determine how groupings of clients value product offerings to ensure it is priced appropriately.
• Subscription-based pricing and product offerings that allow customers to access a bundle of service offerings for one simplified and transparent price (see Lever 3.3 below for further detail).

Lever 3.2 - Explore ecosystem opportunities that increase share of wallet and effectively reach new customers

As productivity has taken centre stage on the transformation agenda, ecosystem capabilities are often perceived as a competing priority. However, in our experience, we see ecosystem capabilities as a strong enabler to unlocking productivity opportunities. 

An ecosystem offering brings multiple providers together through integrated digital platforms, ultimately to provide complementary products that enhance the customer experience. SMEs are looking for a financial institution to partner with them in all aspects of their business, such as lending, cash management/payments, operating their business (e.g., accounting, payroll, taxes) and growing their business (e.g., digital marketing). By participating in an ecosystem, institutions can address a broader range of SME needs, more efficiently attract sticky operating deposits, and unlock growth by using new data to proactively offer incremental solutions.

Lever 3.3 – Drive growth through subscription-based product and service offerings

Just as consumers have grown comfortable with subscription models, so too business banking clients want the ability to add or remove products and services quickly. The EY Global SME survey (2021) revealed that ~43% of Canadian SMEs surveyed expressed interest in subscription-based financial services.

Unlike traditional transaction-based product offerings, a subscription model provides customized offerings that are composed of products for a set fee. Subscription offerings align to the client's baseline needs — including corporate loans, lines of credit, deposit accounts and money movement — and promote add-ons and enhancements — such as payroll or multicurrency pooling — which are available a la carte for an additional charge.

By moving from a transaction-based to a subscription model, institutions can drive profitable growth from:

• Increasing client adoption through a clearly defined value proposition and transparency around fees.
• Making services available to smaller clients that were once reserved for the largest corporations.
• Increasing cross-sell of traditional banking products and beyond banking services, as clients can easily add services to their subscriptions.

The most effective productivity agendas include both cost and revenue initiatives, with short-term cost takeout used to capture quick results and longer-term initiatives for enabling sustained change.

In slower economic times, financial institutions focus on productivity and generally short-term cost takeout initiatives. In our experience, institutions that most effectively improve productivity do so by implementing cost takeout initiatives and thinking about longer-term initiatives that deliver sustained change. This combination enables institutions to meet short-term objectives and shareholder obligations while also moving towards a future that is streamlined and more data driven, all of which enhances productivity over the longer term.

Institutions can start by focusing on quick wins that realize early savings and can fund longer-term productivity work, eventually leading to a self-funded program. When economic conditions improve and institutions revert to favouring a growth agenda, having invested in longer-term productivity initiatives will have prepared the institution to capture growth more efficiently and be more resilient to future economic slowdowns.