Home » Nova »

Incremental Intergration between Business Analytics and Strategy

 

The level of integration between Business Analytics and Strategy depends on the organization’s strategy, internal competencies, technological options, and competitive situation.
We present four scenarios that illustrate different degrees of integration between the business analytics function and the company’s strategy. The purpose of these scenarios is to consider where any organization is in relation to these scenarios. The scenarios can also give some insights into whether the organization has understood and achieved the full potential inherent in business analytics, and thus whether more effort should go into optimizing and maturing the deployment of business analytics.

You will be able to assess the level of integration in your own organization and decide whether the actual level corresponds with the strategic level as well as whether the strategic level uses the full potential of information as a strategic resource.

Scenario 1: No interlink between Strategy and business analytics

In this case, there’s no interlink between the deployment of business analytics and strategy. The most common explanation is that, when developing their strategy, companies often focus on the most visible aspects, such as sales targets, production targets, or cost targets in connection with procurement. In relation with the achievement of these targets, the sales department, production, and procurement will be faced with targets based on the business strategy. However, a company consists of many other functions such as HR, finance, product development, strategy, competitive analysis, administration, and business analytics. These functions are called support functions because they are not adding value in connection with the daily production. However, if they did not exist, the company would encounter problems in the long run.

From a strategic perspective the supporting functions are of course expected to support the primary and value creating processes. This may happen when the support functions themselves interpret the business strategy in their daily activities, but more often by the owners of the primary processes placing demands on services from the support functions based on their targets. So when we describe a scenario with no link between strategy and business analytics, it is not a question of completely uncoordinated entities, but rather a case of a filter existing between them.

A filter may exist because it is primarily the individual processes owners on an ad hoc basis and not the strategy that defines which information is to be generated by the business analytics function.
The consequence of the filter is that the business analytics function prioritizes its tasks according to what best serves the daily target achievement of the company instead of what is best for long-term strategic projects. Moreover, the business analytics function tasks are performed based on the driving force of different users requesting information. In terms of reporting, this will result in the development of more or less authorized reports with inconsistent presentations of the business that they are describing. All in all, the quality of business analytics in this type of organization will typically be an assessment of how quickly a question is answered and how well- founded the answer is.

Other reasons why there is no formalized link between the strategy and the business analytics functions may be that the right conditions simply do not exist. There are situations, such as small businesses with one or few customers, where the cost of running a data warehouse is bigger than the value of the decision support created. And there are companies that define their strategic targets in a way that is not measurable. If, for instance, a company defines a target to be ‘‘we need to establish better relations with our suppliers,’’ then that may be difficult to quantify. Because this definition does not tell us what to measure, we must ask: Is it the number of complaints, average time per transaction, or the quality of their deliveries that are to be improved via this new strategy?

Scenario 2: Functional interlink between business analytics and Strategy

The second scenario represents what we call an adapted information strategy. Here the business analytics function is a reactive element, solely employed in connection with the monitoring of whether the defined targets of the strategy are achieved.
The recipients of key performance indicators (KPIs) are the individual departments, which means that there is no feedback to the strategic level. The business analytics function supports company performance proactively, but only reactively in terms of company strategy. There may be a formalized dialogue between individual functions and business analytics, but the relation to the strategy function is formalized as a monologue, from strategy to business analytics function.

In terms of the quality of business analytics in such an organization, it’s important to be good at defining targets based on strategy. These are targets that relate to each other internally and that, combined, make up a whole. It is equally important that the business analytics function is technically competent when it comes to operationalizing these targets via reports and making those reports both accessible to users and also full of the most updated information possible.

Based on a strategy development process, individual departments define a number of specific requirements, or targets, they are to achieve. Sometimes a target will simply be given to the sales department: It must increase revenue by 10% over the previous year. Alternatively, the department may be given additional information about which segments to grow and with which products. There may also be a message that this must be brought about in cooperation with other departments, such as marketing. Based on the given targets, it will then be up to the individual functions—with various degrees of autonomy—to decide how they are going to achieve these targets.

In this case, the requirement from the company’s overall strategy could be to reduce absence due to illness by 10%. How to achieve this will not necessarily be specified. Consequently, the HR department itself will have to come up with an HR strategy that specifies how it intends to meet the target and a deadline for its achievement given by the company’s strategy. In the same way, a substrategy needs to be developed for customer relationship management (CRM), the CRM strategy, and a substrategy for the production department, the production strategy.

Corporate Targets Requirements
It is important to determine what information may be relevant to the company when developing its strategy and when monitoring whether this strategy is being achieved. In connection with the monitoring of the strategy, a number of targets to be achieved are outlined. These are basic requirements to ensure that measurements can be operationalized. For when is it that we must expect to see the increase in the number of our customers by 20%? How do we define our customers? If, halfway through the year, we see an increase of our customer base of only 8%, is that then a problem or to be expected? We also need to ensure that if we find targets that look as if they will not be achieved, we can actually pinpoint the person responsible, who can then react to this information. Five updated target requirements follow. Some of the requirements are necessary for the technical establishment of benchmarks; others are concerned with who must take action if the benchmarks deviate in a critical way from the specified targets. The requirements we make to benchmarks are that they must be:

• Specific: Targets, such as how many customers we must have by the end of the year, what our revenue must be, by how much we must reduce delivery times, and so forth.
• Measurable: If it’s not measurable, it’s not a relevant target. If we do not know how many customers we have, we need to find another target. If it’s not possible to allocate revenue and costs to the processes we want to improve, we need to establish some other targets.
• Agreed: The organization must accept the targets. If this is not the case, there is no ownership and the organization is about to implement a strategy that, at best, will be ignored or, at worst, will be counteracted. It is implicit, too, that accepted targets mean that we have some specific individuals who are directly responsible for the given targets.
• Realistic: Targets must be realistic. Often, targets are accepted without standing a chance of being achieved. This may have something to do with the corporate culture, maybe someone is trying to buy time, or that there are no consequences involved in not achieving the targets.
• Time-bound: What is the deadline for reducing costs to a certain level and raising customer loyalty up a level? It’s also important that we are able, at an early stage, to determine that targets are not being achieved as expected in order that we can make corrections.

Seen in a business analytics context, objectives need to be specific, measurable, and time-bound so that they can be defined and operationalized in the first place. If they are not, we won’t know, when implementing the technical solution, which information to collect and calculate so that it describes the overall objective of the desired process. If the objectives are not measurable, we cannot quantify them technically and thus measure them on an ongoing basis. Likewise, objectives need to be time-bound if an information system is to be able to deliver messages to users, when critical values are exceeded.

In a broader business context, the five requirements work to ensure a clear-cut understanding of the basis of business initiatives. If objectives are not specific, they may be interpreted differently, which leads to different versions of the truth. If objectives are not measurable, people will start debating whether customers are loyal enough.
For technical reasons, too, it’s essential that benchmarks be time- specific, since the entire establishment of a data warehouse is about relating pieces of information and creating a historic view. If we are to deliver efficient reporting, the time dimension must be clearly defined.
We prefer to automate the measuring via a data warehouse, so that users on a continuous basis can retrieve data about the achievement of targets. However, customer information, such as ‘brand awareness,’’ will typically be distributed through reports. Consequently, we do not need to be able to retrieve all objectives from a data warehouse, but it’s generally preferred, because it means that there will only ever be one version of the truth, and that this truth can be delivered whenever users so desire and in an aggregated form.

Scenario 3: Feedback between the Strategy and the business analytics Functions

The third scenario is based on the existence of an established data warehouse to integrate and store data, as well as an established business analytics function with analytical competencies to make use of this data. We are typically looking at a significant investment in software and employees. This scenario is also characterized by a continuous cooperation between the strategy and the business analytics functions. The reporting methods used at this level for the managing and measuring of operational processes now begin to have different names such as business performance management (BPM) systems, scorecards, and customer profitability/segment analyses. This signifies that a flow of information is going back to the strategy function based on the created reports.

The information described is feedback information from scorecards and BPM solutions. These types of solutions are normally cyclical and start with a strategy. Based on the strategy, three things occur: benchmarking is carried out; there is then an ongoing measuring and analyzing of deviation from targets; and finally, based on the analyses, the strategy is adapted and optimized. We will take a closer look at this in this chapter’s section ‘‘Corporate Strategy’s Subsequent Requirements to business analytics.’’ Quality for the business analytics function in this scenario is the ability to deliver relevant information to the strategy function. This is done in order that the strategy may be adapted on an ongoing basis for the organization to accommodate changes in the market and within the organization itself.

When reports are produced describing whether individual departments are meeting their KPIs, action will, of course, be taken if any major deviations between targets and the achievements are shown. There will therefore always be some form of feedback between target achievement and strategy, although this feedback process may be more or less formalized. An example of the conceptualization of the feedback processes is found in corporate performance management (CPM) and score carding. this is an ongoing cycle, where the enterprise as its starting point has defined a strategy to be implemented in the various departments that make up the business. Coordination is performed by identifying the so-called critical success factors, which are the elements that are essential to whether the strategy is successful, and making sure that they are coordinated. This is typically done via internal meetings across functions. It is at this stage, too, that we define who is responsible for the various KPIs, and thus who must react to these and in what way.

When the strategy is set in motion, progress is measured on an ongoing basis. Generally speaking, KPIs are rarely hit accurately, but rather a bit over or under. In both cases, learning can be derived based on analyses. Did we overlook any potential opportunities, or do we perhaps lack certain competencies in the organization? An optimization of the strategy takes place when we use this learning to improve our business processes and thereby ensure that the organization maintains its agility between the annual strategy processes. And experiences from previous strategy iterations can contribute to create learning in terms of the strategy for the coming iteration.

An alternative way of operationalizing strategic feedback processes is via the ‘‘balanced scorecard,’’ a method introduced in the early 1990s. It connects corporate strategy with the internal processes that will be realizing them, connects it with customer loyalty and, finally— and this was the new thing—with the organization’s internal competencies. What the balanced scorecard achieves, therefore, is to link the primary production processes to the development of the business. If we cannot produce enough, do we then employ more people or different people, or do we establish a dialogue with our employees and on that basis reward them differently? The method, which was developed by Kaplan and Norton, represents a cornerstone for how to formulate requirements in connection with the implementation of a new strategy.

Scenario 4: Analytics as a Strategic Resource
The fourth scenario is about information being regarded as a strategic resource. Such enterprises are characterized by using their analyses of market strengths and weaknesses by systematically thinking about how this information, combined with their strategies, can give them a competitive advantage.

Technical solutions and more about people competencies that are required in the strategy development process. In some cases, this may mean that the enterprise needs to ensure it has staff with both strategic and information knowledge represented at the top management level. This is not an altogether surprising conclusion, considering that we live in the age of information.

A typical example of an enterprise that focuses on information as a strategic resource is Amazon.com, which sells books via the Internet. Here information is saved about the individual customer’s purchases and requests, and these are then processed, with the result that customers are subsequently greeted with offers that are relevant and of service to them. This is a case of improving the relevance of offers to customers based on information, which differentiates Amazon.com in a positive way from other Internet-based book shops. This trend is emerging among certain retail chains, too, where the segmentation of customers means that services can be customized to local conditions.
Moreover, we see a growing sale of information from shops to the manufacturers of the goods sold in the shops. This information describes which types of people buy their products, how price-sensitive the products are, which products are typically sold together, and so forth. This feedback constitutes essential information for manufacturers in terms of product development, pricing, and promotion in the right places with the right messages.

You can distinguish strategies created by a company that uses information or data as a strategic asset by looking at certain elements of its strategy. If a company does not use information as a strategic asset, it will not, in the strategic implementation plans, have descriptions of how the competitive advantages should be gained via the use of information. If a company does use information as a strategic asset, then next to the objectives of the strategy it will also provide directions of how the objectives should be reached via the use of information.

You can also recognize an organization that uses information as a strategic asset on its culture, where the employees, according to our research, intuitively will think proactively in terms of how they can use information to overcome, for example, a new competitive situation. This sort of a culture will use the information as a strategic asset as a result of a top-down process as well as a bottom-up process. This means that if one region learns to improve its processes via the use of information, the news will be captured by the strategy team, and spread as a best practice to the rest of the organization as a result of the next strategy creation process.

Information at a strategic level must be understood centrally in connection with the strategy development process and throughout the organization where the implementation is carried out. Nonprofit organizations need to consider how information is used and regard it as a central leverage in terms of the performance of tasks, which are defined in the organization’s objectives and strategy. As previously discussed, the use of information as a strategic resource is first about identifying central competitive parameters, and second about understanding how this information can ensure that the enterprise differentiates itself from its competitors. We have chosen to introduce a tool that will help you determine which information is going to support your organization’s business-critical initiatives, as well as provide you with a number of examples of how this works in specific terms.

Product Innovation Analytics

The business analytics function needs to deliver information to the strategy development process. Naturally, it’s not always possible to know exactly and at all times which information to deliver. However, based on an analysis of which competitive parameters an enterprise experiences in the market, we can tell which type of information to focus on. That is not to say that any information can be ruled out as irrelevant in advance, but in a world of constant decisions on focus and the prioritization of tasks due to limited resources, some information will have an obvious priority.

This perspective is highly prioritized by enterprises that act based on ‘‘product innovation’’ or ‘‘product leadership’’ strategies, that is, businesses that adopt as their central competitive parameter that their products and services can be characterized as ‘‘state-of-the-art.’’

The knowledge these businesses normally use to differentiate themselves in the market is therefore closely linked to technological knowledge about product development and, to a lesser extent, to knowledge about customer behavior (customer intimacy strategy), or knowledge about how to efficiently produce and deliver services and products (operational excellence strategy).
In the market, analyses are prepared by enterprises on a continuous basis, and models are developed to describe the state of the current market, as well as how this market can be expected to change in the future. Combined with forecasting models for individual products, this can give some useful estimates of where revenue will be earned in the future. These analyses should be supported by the historical knowledge, too, which we stored from the life cycles of comparable products, information that can be expected to be delivered by a data warehouse.

A further dimension of the analytical process is that it is possible, based on the data warehouse information, to examine the development of customer segments. One potential analysis would be looking at which customer segments buy which products, as well as the development of these segments. This analysis can then be broken down on countries, chains, brands, and so on. Often businesses do not just sell one product or service per customer. Some software packages consist of a number of optional modules. Another example is car sales. Special rims cost extra, and the customer can choose not to buy them. Mobile subscriptions typically offer a large number of optional products, a guarantee never to have to pay more than a certain amount per month, favorite numbers that you can ring for less, insurance for the phone, and so forth. The analyst must not only compute the profitability of individual products, but also include any cross-sales related to the product.

If the assumption is that you are serving only one customer segment, you have several techniques to uncover customers’ multiple-purchase patterns. Correlation analysis can use statistics to prove whether there are, between any two services on offer, any positive or negative purchasing tendencies, and thereby show whether the products are complementary or substituting. Principal component analyses(also known as PCA or explorative factor analysis) are able to provide information about how many multiple-purchase patterns there are as well as describe them. We are not looking only at whether products are sold in twos. A multiplepurchase pattern can include five products that are sold together.

If we are working with large numbers of different products , data mining techniques will start to become the most interesting ones. The methods are called everything from basket analysis to cross-sales analysis. The technique is based on uncovering which products are sold together. The underlying methodology is based on the same mind-set as a correlation analysis and involves a simple counting of how often two products are seen together as a multiple purchase on the same receipt or under the same customer number in the data warehouse.

The knowledge these analyses bring to the strategic planning process is the basic information that describes which products secure earnings (not revenue), and where investment in technology seems to be rendering a positive return estimated over the entire lifetime of the product.
When we begin to relate products to each other, knowledge is created about which basic needs we meet for our customers. Products are positioned in relation to each other, in connection with future marketing initiatives. If it’s a matter of whether products can be integrated into each other, such as phones with cameras, GPS units and entertainment, or software packages such as Microsoft Office (spreadsheets, presentation programs, and word-processing programs), then these analyses also provide us with input for future product development.

Furthermore, these multiple-purchase patterns tell us something about our customers’ needs and, as such, provide us with a basis for segmentation based on needs. Are our customers, for instance, buying large aluminum rims for their cars, or are they buying safety such as airbags and fire extinguishing equipment? The additional purchases meet different needs, which means that we should not be using the same sales strategies for all customers. So if a business is serving several segments, a cluster analysis is an obvious choice. Cluster analyses serve two purposes, first to identify how many segments a business is serving, and second to identify their characteristics. Whereas before it would make sense to discuss whether a multiple-purchase analysis is a natural part of strategy development, a cluster analysis would constitute a natural part of the identification and description of the customer groups a business is serving.

Once the segments are identified, we can compare their historic development with analyses of the future, and come up with estimates of what our customer mix will look like in the years ahead. Moreover, we will be able to estimate earnings from the different segments and in this way prepare strategic plans for who we want to serve in the future—and with which products and services.

 

Related Posts

  • No Related Posts