Great Analytics start with a Great Question
Analytics must deliver business outcomes - otherwise it s just reporting 1
In the last five years, many Australian companies have invested heavily in their analytics capability. But is it paying off? Over the course of Accenture s numerous engagements, we have seen how the comprehensive application of analytics can drive operational efficiency benefits, typically in the range of 15 percent to 30 percent of net operating income. Coupled with an abundance of new data sources, from social media to mobile devices, organisations have reasonably believed that analytics would lead to better business decisions that are based on factbased conclusions and quantifiable customer insights. Analytics can do this but only if done right. Assuming your organisation already has the technology in place, as many do, the key to successful analytics is to begin with a great question, and find the right answer before your competitors do. This point of view proposes that to come up with a great question, you should establish a cross-functional team with representatives from analytics, marketing, segment, product and sales. This will guide the analysis efforts to focus on questions with actionable answers that align with and drive core business strategies. Businesses should support their teams with agile delivery of analytics to enable quick, iterative and continually improving capabilities and results. So ask yourself: is your organisation using the right analytics to help you progress from issue to outcome or are you just doing fancy reporting? 2
Analytics needs a starting point Analytics teams across Australia examine and dissect data every day, sifting through past performance and examining customer profiles in search of trends. Sound familiar? You re not alone. Many organisations wait for their analytics team to come to them with the answers, without first considering the question. A common pitfall of analytics, especially when data sources are vast and varied, is to analyse them without the eventual business outcome in mind. Instead, organisations often make the mistake of simply retrofitting scattergun findings back to business issues. While most organisations have invested in the right technology platforms to enable analytics, the lack of strategic intent behind the analytics means that their return on investment can be slow. The risk in approaching analytics without having a clear goal is that too much time is spent analysing and too little time spent answering critical questions. The solution is to start your analytics with a great question a question generated from a strong business hypothesis, with the goal of solving an issue or satisfying a need. Questions need to come from a variety of sources beyond just your analytics team. A great question must be linked to a business outcome, and there must be a compelling reason for asking the question the so what factor. Don t focus on finding a perfect answer. Instead, focus on knowing more than you knew before, putting insights into action and producing value. For example, when a large Australian client wanted to develop new and innovative offers for their customers, they did things the way they had always done them: Let s ask the analytics team to resegment the customer base and identify target groups of customers with similar needs and behaviours. In addition, we should ask them to build a new propensity model so we know who is statistically more likely to take up an offer, or who is about to leave us. The Accenture team saw potential value for their client in challenging the status quo. Rather than spending 12 weeks building the new segmentation and propensity models, the Accenture team suggested value could be achieved faster by taking a businessled approach. Working with the sales, product, marketing, segment and analytics teams, the Accenture team developed more than 20 ideas for new and innovative offers that could be tested using analytics. For example, one idea focused on a loyal group of customers that conducted regular transactions with the large Australian client. The client wondered if these customers were reaching a point in their relationship where they would begin to take small pieces of their business to a competitor. A business question was established: Is a portion of our customer base gradually taking some of their business elsewhere ( silent attrition ), and if so, is this of greater concern to us than complete customer defection? Using analytics, the Accenture team identified a group of customers that had been silently taking small pieces of their business elsewhere. Within three days (rather than 12 weeks), they determined that this gradual loss resulted in greater revenue loss than complete customer defection. A recommendation was made, in collaboration with the wider teams of sales, product, marketing and segment, to develop a proactive offer to stem gradual churn. In this example, the analytics team was still employing the same modeling techniques they normally would. The difference was that they struck a balance between letting the data decide and having a clear business objective. The clear business objective gave focus, helping them deliver value more quickly. 3
Great analytics start with a great question Generating a question is about approaching your highestpriority business problems from a different angle so you can solve them innovatively. This is why it is so important to bring together individuals from your analytics, marketing, segment, product and sales teams to help generate questions. Figure 1. Four steps for successful analytics Measure 4. Scale for business outcome 1. Define the question: Bring your teams together to assess the most important business problems for analytics to answer. 2. Perform analytics: Using the question to guide your thinking, run analytics to find your answers. 3. Test and learn: Validate your insight by applying the findings to a small sample. 1. Define the Question Business buy-in Sales MeProduct Analytics Segment & Marketing 4. Scale: If you got it right, apply the insights from your analytics across the full extent of the initial business issue. asure and Go/No-go 3. Test and Learn Go/No-go 2. Perform Analytics As shown in Figure 1, there are four key steps for successful analytics, and the first one is all about defining the right question. To define the right question, organisations should: Create new ways of doing things. In this model, analytics questions are not generated in a vacuum by the analytics team. Instead, it requires buy-in and participation from multiple teams, including product, analytics, marketing, segment and sales. Encourage maximum collaboration between all stakeholders to generate questions. This requires individuals from product, analytics, marketing, segment and sales teams to collaborate, test and prioritise questions accordingly. Leverage existing forums where appropriate, to enhance output and cultivate collaboration. Link to business outcomes. Adopt an approach that links your questions and their so what to how they would help drive the organisation s strategic imperatives or existing tactical initiatives. Always confirm stakeholder buy-in to a targeted question. Does the question really need to be answered? Will the answer deliver critical business intelligence that helps further a business goal? Think ahead to how you ll answer the question. When generating your questions, consider the obstacles you might face in answering them and start to address them. These actions will help you get the most out of your team s experience and create ownership for people s questions. More importantly, it will help your organisation build a strong platform from which to perform your analytics, test and learn, and then scale the opportunity across the business. We ll discuss these steps later. 4
How to come up with a great question Measure 4. Scale for business outcome Sales Analytics 1. Define the Question Business buy-in In creating a great question Product Segment & Marketing Measure and Go/No-go 3. Test and Learn Go/No-go 2. Perform Analytics the role of sales is to: Take a lead role in the development of the new questions and campaigns, with support from the marketing, segment, analytics and product teams. Sales are integral in identifying opportunities for further analysis (based on direct contact with customers) and helping provide real-time feedback on the products and offers from the customer s perspective. the role of segment and marketing is to: Identify how the question will apply to the customer segments within the target market. When coming up with questions, marketing and segment represent the customer from the organisation s point of view. Segment should provide a perspective on the opportunity, competitor activity and areas for exploitation. Marketing will guide the question to answer how channels, event-based triggers and offers can be used to most effectively flag and target specific customers. Together, they must also address any sales or service improvement opportunities that come out of the analytics. the role of product is to: Help to define the question, focusing on offers, features and pricing. Product is critical to providing a perspective on the cost and benefit of any changes to product dimensions. Product uses analytical information to drive their decisions, and should be embedded in the process from the beginning. the role of analytics is to: Assess the validity of new questions by looking at the size of the opportunity and expected take-up rate, based on what has worked in the past. Analytics bring their aptitude and experience in developing insights from analysing data in new ways. The focus for the analytics team includes delivering insights that assist in forming new offers (for example, increasing revenue or reducing cost), performing additional data analysis (as agreed upon by all stakeholders), and championing a culture of data-driven decision making. Once the question is defined, they will need to develop the contact strategy and creative requirements of the analytics-driven campaign, or tailor the organisation s channel and service experience to address the findings. 5
Linking analytics to business outcomes makes a powerful business case Have you ever tried to develop a business case for improving your organisation s analytics capability? When we see organisations struggling to quantify the benefit associated with analytics, it is often because they are not involving all parts of the business. Picture the scenario where your data owner, usually IT, wants to put together a business case to improve analytics. Analytics is the right thing to be doing, but from IT s perspective, the business case will be more about cost savings than revenue generation. To add more of the so what and justify the time and expense of developing a robust analytics capability, the business case needs a revenue lens. This is what a customer and sales view can bring, making the business case truly compelling. A combined IT and business perspective can help put together a much stronger, more persuasive argument for building the reach of analytics within your organisation. Figure 2. To be truly compelling, a business case for analytics requires both IT and business Analytics must support business initiatives. It is not an independent endeavour. Business Business Strategy Business Initiatives Business Outcomes IT Information Strategy Information Management Information Analysis 6
Examples of great questions to be answered with analytics (Tied to a business so what ) Measure 4. Scale for business outcome Measure and Go/No-go 3. Test and Learn Sales Product Analytics Segment & Marketing 1. Define the Question Business buy-in Go/No-go 2. Perform Analytics Customer Service We offer a range of sales and service channels. Are there patterns in a customer s recent service interactions across these channels that would help us to predict a future purchase or indicate an early warning signal for possible churn? If we know this, then so what? This insight can help us target customers who are likely to purchase, or those that are at risk of churn. For example, analytics may detect a significant increase in customer enquiries that could go either way: they are looking to purchase a new product, or they are looking to leave us, as indicated by their comparison shopping behaviour. Sales Staff Performance within our sales team varies significantly. What are the characteristics shared by our top sales people that are key to their success? If we know this, then so what? If we can understand the sales competencies, personality traits and behaviours of high performers, we can see the path to shifting and reshaping the performance curve of our sales force. Products We believe groups of similar customers buy or use our products in a similar way. For example, that they use the same channels, purchase in the same sequence and repeat purchases at regular intervals. Is this true? If we know this, then so what? If we understand a customer segment s purchase behaviour today, we can logically offer customers with similar characteristics the most obvious, next-best product, thus driving increased sales. Processes Which common customer call centre transactions take the longest? For these transactions, which processes do employees follow, and which tools and resources do they use? If we know this, then so what? For our key processes, we can identify errors, bottlenecks and training gaps to improve productivity and customer experience or even reduce the likelihood of a customer triggering a particular process to begin with. Competition In the current aggressive and competitive market, it will be hard to win new business. How much scope is there to leverage our existing customers to stay ahead of our competitors (for example, with increased shareof-wallet or reward referrals for new customers)? If we know this, then so what? Increasing sales within the current customer base and relying on a referral strategy to win new business may not drive the levels of growth required. We may also need to remain competitive or be aggressive to win new business. New revenue streams Are other companies interested in our data, thereby creating a possible additional revenue stream for us? If we know this, then so what? Our customer base and the information we hold on them (within privacy legislation) may represent a new, incremental revenue stream. Information can be packaged and sold to new and potential customers, resulting in game-changing product innovation and partnering. This approach is commonly referred to as data monetisation. 7
Answer the question then act on the answer Answering your analytics question may seem like an obvious step, but it s one that many organisations struggle with. The key is not in perfecting all aspects of the question to derive an exact answer. Rather, focus on finding an answer that s more accurate or relevant than you had before. This avoids analysis paralysis. Measure 4. Scale for business outcome Measure and Go/No-go 3. Test and Learn Sales Product Analytics Segment & Marketing 1. Define the Question Business buy-in Go/No-go 2. Perform Analytics The remaining steps in the model are as pivotal to the analytics function as defining the great question, but are not the focus of this point of view. Thus, we will examine them briefly. Analytics will answer: What is it today? What could it potentially be? What is the value of the difference? If analytics shows that an opportunity exists (that is, the potential value is greater than the value today), then the next step should be to take action. In the case of the European bank ABN Amro, analytics were first employed to gain a better understanding of how customers were using their website. The key question would have been: How can we optimise the website to promote use and improve the customer experience? For ABN Amro, the steps highlighted on the previous pages were key to distilling their findings into actionable insights. This meant using web metrics to identify issues with the website, so as to improve its structure, layout and content. The so what was to initiate and measure test pages that highlighted a stronger and more prominently placed call to action. By making these changes, ABN Amro was able to improve their web conversion rates by 290 percent thirty times the uplift they had anticipated. Once you have an answer, the next step is to relate it back to the original core business question. As in the case of the large Australian client highlighted previously (see page 3), the hypothesis is an important foundation for the next steps of building an analytics strategy and approach. In other words, to help ensure that your insights are valuable, you need to establish a scientific approach for acting on the answer. Example of acting on an answer: A marketing campaign A test-and-learn approach, in most cases, should be used if the action on the answer is to target a group of customers via a marketing campaign. For example, if the organisation questioned whether they had a source of untapped customer potential, and the analytics validated the hypothesis that a large opportunity did exist, they could develop an offer to appeal to this potential customer group. Prior to delivering a fully scaled and multi-channel campaign, the organisation should test the offer with a small group of customers. After testing the campaign, the learnings are used to adjust the next iteration of the campaign. This cycle may repeat a number of times. Only when there are no more adjustments required, and the test has shown itself to be successful (that is, when test campaign response rates are above control or hurdle rates) should the campaign be scaled. This approach requires descriptive progress tracking and review activities to ensure continual learning and improvement. The goal here is to rapidly enhance the quality and relevance of the analytics insight given to sales staff. A successful test-and-learn approach will want to achieve the following outcomes: Refine and optimise the campaign, as well as iron out any problems prior to scaling. Create a more responsive process for repeating successful campaigns. Focus the marketing effort towards quality and applicability, rather than volume. 8
Measure the effectiveness of your answer Measure 4. Scale for business outcome Sales Analytics 1. Define the Question Business buy-in Measure and Go/No-go Product Segment & Marketing Things change over time, but few organisations revisit previously established assumptions until they are well past their effective expiration date. It s important to ensure that your answer is still valid and to adjust it if required. Here are three steps to make sure: 1. Always link your question to measurable outcomes, such as increased revenue, reduced costs, improved customer satisfaction or better staff engagement. 3. Finally, use return on investment (ROI) hurdle rates to validate the financial feasibility of acting on the answer. Did we get it right? By following these three steps, you would continually learn more about your question, your answer, and most importantly, your customers and what works. 3. Test and Learn Go/No-go 2. Perform Analytics 2. Measure the impact of your answer. For example, if your organisation generated an offer to exploit a new customer market, understand the success criteria and measure the financial outcomes. 9
High Performance. Delivered. Remember, it s not about finding the perfect answer. It s about knowing more than you knew before, and putting the insights into action to deliver business outcomes. Gaining a competitive advantage with analytics requires more than the best analytics technology and the smartest people staring at data all day. It requires a great question as a starting point: a strong, business-driven question to guide and focus the organisation s energy and resources. This question should be developed with a range of stakeholders that each brings a valuable customer perspective to the organisation s key challenges. The organisations that are the most effective and efficient at answering their great questions, then taking and measuring action, will stay ahead of their competitors and lead the pack. 10
Contacts Philippe Konfino Managing Director S&CS, Asia Pacific Philippe.Konfino@accenture.com Alex Burrows Senior Manager, S&CS Australia & New Zealand Alex.Burrows@accenture.com About Accenture Accenture is a global management consulting, technology services and outsourcing company, with 257,000 people serving clients in more than 120 countries. Combining unparalleled experience, comprehensive capabilities across all industries and business functions, and extensive research on the world s most successful companies, Accenture collaborates with clients to help them become high-performance businesses and governments. The company generated net revenues of US$27.9 billion for the fiscal year ended Aug. 31, 2012. Its home page is www.accenture.com. About Sales & Customer Service Accenture s Sales & Customer Service (S&CS) service line helps organisations achieve high performance by transforming their marketing, sales and customer service functions to support accelerated growth, increased profitability and greater operating efficiency. Our research, insight and innovation, global reach and delivery experience have made us a worldwide leader, serving thousands of clients every year, including most Fortune 100 companies, across virtually all industries. Disclaimer This point of view is intended as a general guide and not as a substitute for detailed advice. Neither should it be taken as providing technical or other professional advice on any of the topics covered. So far as Accenture is aware the information it contains is correct and accurate but no responsibility is accepted for any inaccuracy and error or any action taken in reliance on this publication. This publication contains Accenture copyrighted material and no part of it can be copied or otherwise disseminated with Accenture s prior written consent in each case. Copyright 2012 Accenture All rights reserved. Accenture, its logo, and High Performance Delivered are trademarks of Accenture. 12-1689