This chapter is designed to explain the profile of a data analyst, their role within a strategic organisation and the knowledge and skills needed by an analyst to contribute positively to the challenges of the analytical business environment. This chapter particularly focuses on the soft skills and competencies of an analyst’s role.

WHAT IS A DATA ANALYST?

A data analyst is someone who analyses data to find patterns, trends or hidden information and translates these into insights that can be useful to business. It is a critical role in modern organisations.

Data analysts make sense of the noise that exists in data: they aggregate and translate this data into relevant business metrics and analyse it to provide meaningful insights relevant to their organisation’s decision-making needs.

A successful analyst becomes integral to strategic decision-making and can grow his or her career into critical business roles within business intelligence and strategic planning. Analysts can be from a variety of educational backgrounds. While there is no mandatory qualification required to become a data analyst, a degree in computing, mathematics, statistics, economics or research is helpful. Analysts also benefit from education and experience in fields as diverse as sociology, humanities, marketing, engineering or other research degrees.

AFFILIATED ROLES AND DIFFERENCES

Data storage architecture and warehousing became popular in the 1980s. Early on, the focus was on storing data efficiently only for the purpose of producing basic management reporting. Since then, data analysis has come a long way and there are now various roles that support data-driven initiatives in an organisation. Roles have been created as the use of data has evolved and the complexity of managing and leveraging data has increased.

Nearly all data-related roles have some overlap; some of the key data roles are shown and described in Figure 2.1. Data analysts should collaborate and work with these other roles.

Figure 2.1 Key data-related roles and differences

KEY INDUSTRIES WHERE DATA ANALYSTS WORK

The role of data analyst is not limited to any particular industry. However, some of the key industries where data analysts thrive are:

Industry-specific knowledge requirements

There are not necessarily any specific industry knowledge requirements that hold a data analyst back from taking up a role in an industry where the analyst has not worked previously. It may be beneficial for the analyst’s job prospects to have experience of working in the industry where the vacancy exists, but this is not always mandatory.

To be successful in the job, a data analyst needs a mix of three main types of skills: functional, technical and soft (see Figure 2.2). These are explained in detail later in this chapter, with a particular focus on the role of soft skills.

If an analyst is able to migrate functional and technical skills across different industries, and invest in continually developing these along with targeted soft skills needed for that organisation, they can easily move between industries.

As an example, an analyst working in insurance with advanced mathematical knowledge can target the IT industry based on existing functional and technical skills, but will need to become familiar with the business challenges particular to IT setups.

Figure 2.2 Skills needed as a data analyst

Certain industries, and specific roles within them, might require certain knowledge or aptitude. A credit risk data analyst, for instance, might be well versed with the regulations relating to Basel norms,¹ but might not have the necessary skills to work with documenting the results of a clinical trial in the pharmaceutical industry. While it is impossible for a candidate to know all the nuances of each industry, a general understanding of how an industry operates is often critical to be able to transition across roles in various industries with ease.

NATURE OF TASKS UNDERTAKEN BY DATA ANALYSTS

A data analyst is usually involved in solving business problems specific to the challenges faced by the industry they are in, or problems emanating from the life cycle that a particular organisation works with. For example:

These are examples of industry level problems that analysts can be engaged in. They may vary widely between organisations within the same industry, depending on the life cycle within that industry.

Let’s explore the various stages of a product’s life cycle within an organisation (see Figure 2.3) and the typical tasks that a data analyst might be undertaking within these life cycle stages.

Figure 2.3 Typical product life cycle

Availability of data and its impact on an analyst’s scope of work

The amount of data available can influence the way data analysts approach problems. Take the examples below, for instance.

Let’s say an organisation currently in the growth stage of a product is trying to deal with the poor app performance that its users are facing. In such a scenario it is likely that there is abundant data available on the usage of the app and the specifics of when the users experience issues. The analyst, in this case, has data to analyse and can help to identify the reason for the problems.

In another scenario, let’s assume the same product is dealing with low usage issues. How can the analyst try and understand the reasons for low usage? If there were multiple segments of users of the app, with some heavy and some light users, the analyst could try to profile both segments to gain insights. In such an instance, the analyst will have to collaborate extensively with the business teams involved and try to check various hypotheses on why the usage is lower than expected. The analyst will also have to check if any of the team members have been receiving qualitative feedback, and could suggest that the business conduct a primary research survey and seek feedback on user experiences if necessary.

Even in a scenario with abundant data, the data analyst may try and incorporate additional sources of information to analyse. If the analyst is analysing sales growth in the past year and trying to make sense of the forecasts generated for the next few quarters, they may try and include data about the macroeconomic parameters published by the central bank.

While designing a new process to screen the eligibility of applicants, a data analyst may suggest that it could be useful to approach one of the leading credit bureaus to seek further information about the applicant and use it to augment the data already held by the company.

DATA ANALYST KEY RESPONSIBILITIES

The responsibilities of a data analyst depend on the industry and organisation they are working within, and are also influenced by the level of investment in data their organisation makes. Broadly, these are the key responsibilities of a data analyst:

Extraction

This usually entails extracting data for reports and projects that the data analyst is tasked with producing. Certain business events may also trigger a request for customised data extraction.

Analysis

The data analyst would then generate insights based on the data and the business understanding. Inputs from subject matter experts (SMEs) may be required at this stage. Some insights from the data may be counter-intuitive to expectations and this could be a result of data quality issues or a gap in business understanding of the problem. SME involvement would help to corroborate the insights generated.

Transformation

The data analyst can then share the insights with stakeholders using a data visualisation tool. Data visualisation entails using tables, graphs or other measures to present the data insights in a clear, visual manner. Seldom would the analyst present insights without some form of data visualisation.

DATA ANALYST KEY SKILLS

As mentioned earlier in this chapter, an analyst needs a blend of three types of skills to be successful: functional, technical and soft skills.

This chapter focuses on soft skills to highlight the importance of these skills in the role of data analyst. When dealing with experts with similar technical skills, the only distinguishing factor among these experts could be their soft skills.

Key functional skills

Key technical skills

Key soft skills

Data analysts need a variety of soft skills in line with the tasks they perform on a daily basis. These include the ability to deal with challenges, stakeholder management, maintaining multiple communication channels and presenting results. Figure 2.4 lists some data analyst key tasks and associated soft skills.

Figure 2.4 Key tasks and soft skills required

Let’s look at these skills in more detail.

Dealing with challenges

Some of the challenges that a data analyst faces emanate from data issues, processes, stakeholders and systems. The analyst can deal with challenges by working on the following soft skills.

Adaptability As the Greek philosopher Heraclitus is reputed to have said: change is the only constant. A data analyst needs to be prepared for change. Whether finding solutions to business problems, or business process engineering to create a new or improved process, a data analyst is constantly involved in change.

A data analyst seldom deals with the same problem over a long period of time. Even if the problem remains the same for some time, the data and its inherent message keep constantly evolving.

As an example, data analysts working for telecom companies over the last couple of decades would have noticed changes in the business environment. From being an item of luxury or convenience, the mobile phone has now become a necessity. Voice calls are not the primary purpose of these devices any more: messaging and internet usage are the predominantly used functions. Data captured in this instance would highlight the change in customer behaviour. Mobile operators have had to adapt by changing the tariffs offered as a result of data analysis of customer behaviour.

Not all change is the outcome of industry changes or customer behaviour. Some change may be forced by regulators. Either way, a data analyst has to be prepared for continual change.

Adjusting to learning curves For a data analyst, the learning never stops.

Data at one point, in the infancy stage of computing and digital storage, was stored on floppy disks and exchanged among various teams and businesses. Not all aspects of a customer’s banking transactions were digitally recorded prior to the advent of core banking systems. These days, every aspect of a customer’s interaction with an organisation is recorded and can be analysed. It is quite normal to hear the voice on a bank’s phone menu say that all calls may be recorded for monitoring or training purposes. Every click on a grocery retailer’s website is recorded to help analyse customer behaviour.

A data analyst needs to constantly learn new skills and adjust to a steep learning curve in the shortest time possible to deal with various changes in the industry. The process of retrieving data, assimilating it, making inferences and sharing the results has undergone a lot of automation in recent years and has also been the subject of much analytical research.

A data analyst needs to be aware of the technologies intended to augment the daily job. While it is impossible to keep pace with all advances in the industry simultaneously, having the ability to adjust quickly to steep learning curves does no harm in maintaining standards in the workplace.

Ethics While being ethical is in itself not a soft skill, there are various aspects of ethical behaviour at the workplace that can be developed to ensure that high standards are maintained at all times.

Ethical standards are an underlying requirement for almost everything we do in life but in the case of data analysis, these standards come to the forefront. This is because of the sensitivity of data that an analyst has access to. Prior to appointment, every data analyst would have been subjected to some level of reference checks to ensure that they can be trusted with data.

A data analyst needs to be self-aware and have the best interest of stakeholders at heart. Analysts need to understand the risks that their business processes face. They need to have the courage to escalate any issues through the right medium – be it a manager, whistleblowing helpline, law enforcement officer or whatever. Data analysts are also responsible for ensuring data integrity.

A data analyst should not be mistaken for a cybersecurity expert; however, they can still contribute to ensuring that best practices are adhered to and highlight any potential security lapses relating to data storage and access.

Data analysts can make significant contributions in the workplace by bringing a positive work ethic and dealing effectively and efficiently with any mistakes. Some mistakes made by data analysts could have far-reaching consequences. A mistake itself might be a basic error. For example, a data analyst may have utilised the wrong monthly data file to deduce insights about the passenger load in an airline. The management committee’s action based on such a report may lead to significant problems. What if the data analyst realised the mistake after the insights had already been presented? Should they highlight the error? Not only does the analyst need to highlight the error, in many circumstances it is the analyst and the related team that need to produce a new set of figures and related inferences, and quickly. The error might have been caused by some inconsistent messaging from the management team itself; however, even if this is the case, the data analyst needs to raise a hand and highlight the error and then try and help to fix the fallout from the situation.

Stakeholder management

Managing stakeholders can be a difficult task. For example, consider the importance of communication with stakeholders in the production of a brief for a new data analysis project.

Prior to investing any meaningful time to solve a business problem, data analysts need to understand the brief. A well-structured brief contains background, describes the business problem and often mentions an expectation of how the output should look. If stakeholders have not shared a brief, then the data analyst can always put one together and share it with the stakeholders, getting their feedback to ensure that the problem is understood correctly.

A lot of projects fail to deliver when the brief is not shared, is ambiguous in defining the problem and/or the expected solution is not well understood by the data analyst. In certain business situations, the brief might be ambiguous initially, but after some analysis and preliminary insights, the data analyst can work with stakeholders to rewrite it.

This section will cover the key soft skills that data analysts need in order to deal effectively with stakeholders.

Listening This is not the same as hearing, which is a physical attribute. If a data analyst is a good listener, the analyst would accurately understand what is being said and interpret the message correctly.

Listening includes both verbal and non-verbal messaging. At times, the body language of stakeholders is important when the business problem is being described. For example, leaning in can indicate that they are interested and actively listening. Slouching in a chair in a meeting, however, may indicate that they are in passive mode or, worse, uninterested.

Some people are keen to reply rather than listen. By listening, a data analyst is trying to process as much information as possible. In certain business scenarios, the business problem might have already been partially worked on by another set of analysts. In this case, the data analyst needs to understand the progress made by other individuals in attempting to solve the problem.

Listening is also important for an analyst to perform a role within a team. Whether a team leader or team member, a data analyst has to actively listen and take on board multiple views, for instance while conducting focus group discussions to understand customer preferences. A lot of business metrics have also evolved around listening. Net promoter score (NPS)⁴ and voice of customer (VOC)⁵ are examples of such metrics.

Analytical thinking Data analysts must apply analytical thinking and conduct critical reviews. Going back to the example of the project brief mentioned in the introduction to this section, analytical thinking is necessary to understand it fully and refine it if appropriate. Often, stakeholders are unable to give structured requirements or understand if the solution being proposed is achievable. If this is the case, data analysts should come up with a project brief themselves prior to analysis, and be clear with stakeholders what the expected results will be.

Data analysts also need to have a critical mindset when being asked to conduct analysis; they should only be investing time in the analysis that makes sense from a logical point of view. At times, a data analyst is the only expert on the subject being discussed and it becomes imperative that the data analyst guides the stakeholders wisely and works with them to refine the project brief. An effective data analyst also has to learn to say ‘no’ to pieces of work that are impractical for various reasons.

Negotiation skills These come into play as soon as a data analyst is approached with a project brief. The business may want to do a very large piece of work in the shortest time possible. The data analyst has to try to accurately estimate the time it will take to deliver the task and negotiate with the stakeholders agreed start and end dates for the analysis. At times, the negotiation may entail convincing the business to reduce the scope of the work or even sponsor resources for larger pieces of work.

Data analysts might be dependent on various other teams within an organisation to deliver a piece of work. They may have to implement service level agreements (SLAs) with various teams to ensure that there is an understanding of the lag between raising a request and receiving the required information to analyse the business problem.

Conflict resolution Data analysts have to keep working towards solving the business problems at hand while simultaneously dealing with any interpersonal or inter-departmental issues that may arise.

At times, when various departments have differing approaches to solving a problem, or conflicting priorities in the short run, disputes may arise that could affect business deliverables. Not all such conflicts can be done away with completely. Data analysts must work to minimise such conflicts and use team members within the organisation as mediators if necessary. However, the first step to resolve most conflicts should ideally be a face-to-face discussion, which is held after thinking through the issue.

Problem-solving Stakeholders look to data analysts to solve problems. This may mean that the data analyst has to deal with a whole bunch of other problems to try and solve the particular problem troubling the stakeholders.

For example, a stakeholder might require an important extract of data very quickly. A new data analyst within the team, tasked with solving the stakeholder’s problem, may face issues such as getting access to the right data systems, understanding the data structure and making sense of the business processes in the new environment. Given such constraints, the data analyst is still expected to deliver results. In such circumstances, what matters is the problem-solving attitude of the data analyst. Rather than being overwhelmed with issues, the data analyst needs to try and work through each problem systematically.

The importance of saying ‘no’ to a brief when necessary was mentioned earlier. Having said that, however, data analysts should refrain from refusing to undertake a brief just because the problem is complex. At times, the solution may not lie with the data analyst. In such circumstances, the role of SMEs becomes important. Analysts can turn to them for advice and help in solving the problem at hand.

Task management

Seldom would a data analyst work on a single task at any given time. If it is a long, drawn-out project with a single business objective, the project may need various small streams of work in order to deliver the objective; hence, the ability to manage various tasks at once becomes important. In task management, time management and prioritisation skills overlap, because both of these aspects are needed to effectively deliver tasks on time. The data analyst has to determine their investment in a task based on the time available and its priority.

Ideally, tasks should be managed using some form of documentation. For instance, a Gantt chart may be used to display the project schedule and track the progress made; and a Fishbone (also known as cause and effect or Ishikawa) diagram can be used to identify the causes of a problem.

Different data analyst tasks can be viewed collectively as part of a project. When data analysts work as part of a small team, they will need to be able to manage their own projects within the requirements specified; and when they work as part of a larger project, they will need to liaise with professional project managers.

In this section we’ll explore the different skills related to task management.

Time management Data analysts need to estimate the time required for completing each task, both on a project basis and on a daily basis. At times it is not easy to come up with an estimate. In such instances, data analysts can benefit from breaking down the tasks into subtasks. Assigning a time component against each subtask can help to estimate the overall time required. Often, in the middle of a task, a data analyst might realise that the time needed to complete it is longer than the estimation. In this case, it is important to inform the stakeholders concerned, and discuss with them whether to postpone the due date accordingly or provide a simpler version of the analysis. In this case, the analyst would still need to deliver results with a sufficient degree of accuracy and quality insights.

To manage time effectively, data analysts also have to look inwards and ensure that they have a fair degree of understanding about:

Prioritisation Data analysts may have to deal with multiple unexpected and critical requests, so effective prioritisation of tasks is necessary. The prioritisation process should be managed with stakeholders and/or the project manager. An efficient way to prioritise is to be transparent with the stakeholders. Having a published list of tasks helps stakeholders to understand the current status of tasks and any reasons why their particular tasks have not been prioritised or are taking longer than planned. This could save a lot of time and minimise potential conflicts. It is not uncommon for data analysts to host periodic meetings between various stakeholders and seek their help in prioritising tasks. At times, given the urgency of some tasks, certain stakeholders may be willing to de-prioritise their own requirements to help their peers in other teams.

Decision-making Data analysts may be presented with a problem or situation where a decision must be made regarding a process, data or some other work-related aspect. In an ideal situation, the data analyst would seek a consensus among all stakeholders prior to deciding; however, such a luxury might not be feasible, and the analyst might have to make decisions quickly.

At times, data analysts might not have detailed information, but may need to make a decision or put forth a proposal to the stakeholders based on limited information. Any significant independent decision that an analyst makes must be communicated with the stakeholders. The analyst should try and elicit a response from stakeholders for any outstanding issues at the earliest opportunity where the analyst is unable to make an independent decision.

Some of questions to consider before making any decisions are:

Some of the common decisions that data analysts may have to make relate to:

Communication

The complexity of the techniques used for data analysis often creates an atmosphere of mystery around the discipline, and this can create mistrust of the resulting outcomes and recommendations. Many colleagues that data analysts have to collaborate with do not have specific knowledge of the computational and statistical methods used, and data analysts are often required to explain complex ideas to non-specialists.

It can often be a temptation for analytical specialists to use technical jargon and to overemphasise the advanced techniques that they have used. This might create a sense of wonderment and admiration for the analysts among non-technical colleagues, but could also result in a reduction of the critical debate of the consequences of the analytical results.

To meet this challenge, the analytical community must improve their ability to tell the story of their results without resorting to technical jargon. The technical details of the analysis must be reserved for internal debate about the merits of specific methods within the data analysis community.

It is also important to engage with other professionals on their terms, instead of insisting that they engage on ours, to explain what implications the analytical process will have on their working practices.

In this section we will look at communication skills in further detail.

Establishing communication channels Data analysts need to decide which stakeholders need to be involved with the various sets of communications being shared about a data analysis project. Often, multiple communications about the same task may be sent out, although these may differ in granularity and frequency. It is also necessary to establish the best medium of interaction. In an environment where teams are globally spread out and may be multilingual, deciding on the channels and frequency of communication is important. Not all recipients will respond to your communications; most stakeholders are ambushed with a host of communication, and at times it might be difficult for a data analyst to ensure that their messages get through.

Some stakeholders may prefer to take a backseat and be isolated from the process of problem-solving. Other stakeholders may wish to be actively involved for the full duration of analysis to ensure that they have visibility of the process and can contribute when necessary.

Keeping the audience engaged One of the ways to keep the audience engaged on a project is to share its milestones and goals. These must be established, to keep track of achievements, and can help to maintain focus on the overall deliverables of the analysis.

Milestones and goals help to check that the project is on track to deliver as expected. Communication about them assures the stakeholders that key tasks are being completed and serves as an opportunity for everyone to feed back on the process. We mentioned earlier that at times it becomes difficult for data analysts to estimate the overall time required for analysis. By breaking the time into subtasks, or milestones and goals, analysts might be able to estimate the time required for such activities more easily and thereby calculate more accurately the overall time required for analysis. Once the milestones and goals are achieved, they can be communicated to a wider audience.

In some projects that fail to deliver the expected results, there is a lack of appreciation of the effort invested by data analysts in pursuing those results. It is in the interests of data analysts that a hypothesis-driven approach is adopted for analysis. A hypothesis can be defined as a set of beliefs or preconceived notions about the nature of or solution to the problem that a stakeholder has. It is rare to find a business manager who does not have certain beliefs regarding why a business metric performed in a certain way or what typical customer behaviour is. As part of the analysis, data analysts can occasionally put these hypotheses to the test and also put forward their own list of hypotheses that they aim to prove or disprove. By following this approach, even if the end objective of the analysis is not met, data analysts can show the results of hypothesis testing and thereby showcase the effort involved. Stakeholders can also gain some benefit from the exercise after looking at the results of such testing.

Feedback Earlier we mentioned the need for a project brief that describes a task for the analyst and the outcome that is expected. Data analysts should use this document to assess whether all the information, data and tools needed to undertake the task are available to them. This scrutiny will enable the data analyst to provide feedback to the stakeholders, which is important in managing expectations. To be able to provide feedback effectively, data analysts should:

Presentation

While it might have taken days, weeks or months to complete a piece of analysis, data analysts will usually only get a small proportion of that time to present the results. Results may be presented in the form of a Microsoft PowerPoint presentation, a written document or even a set of metrics in a spreadsheet. Presentations could be given mid-project when they relate to milestones achieved, or are otherwise given at the end of a project.

During some presentations, the objective could also be to elicit feedback to incorporate in the remaining phases of analysis. Either way, what is required is the complete attention of the stakeholders. To achieve this, the presenter must understand the audience, articulate their thoughts effectively and aim to raise the team profile.

Understand the audience Data analysts should be aware of the differences in presentation style and content to a technical versus a non-technical audience. For instance, a technical audience consisting of senior management from the insight team may be interested in knowing details about data sources and issues about data quality, whereas executive management would probably be more interested in knowing about the insights from the data and the business impact of this information.

Other instances of differences in presentation style and content include presenting to specific teams such as sales, marketing, IT or HR. At times, analysts might get a mixed audience in a room. Various audience members may be interested in specific aspects of the presentation: a statistical modelling manager, for example, may be interested in the data cleansing and transformation done by the data analyst; whereas a model implementation manager may be keen to understand if the data used is going to be consistently available in the organisation so that the solutions implemented can be run seamlessly; and a stakeholder involved in governance might be looking for assurances about the process followed in maintaining the standards of the analysis.

Some aspects that data analysts should keep in mind about their audience are:

Articulation Articulation relates to the ability to present complex solutions in a simple and concise manner. A well-articulated presentation showcases the clarity of thought and level of expertise that the data analyst possesses.

Data analysts often make assumptions about data and the business scenario. These assumptions need to be articulated and documented prior to sharing the final analysis outcome. The insights should be supported by data and appropriate statistical test results. The use of jargon should be avoided, unless of course the audience is well versed in the acronyms and terminology used – sometimes not using the industry terminology could be misunderstood as a lack of familiarity with key metrics and standards used within an industry.

After sharing results with stakeholders, data analysts need to:

Raising the team profile Data analysts should view a presentation as an opportunity to raise the profile of their team and not just a particular piece of work. Even if the stakeholders are internal, they are in effect the customers of the team and they are consuming the service that the data analysts provide. A satisfied customer will lead to a sustained level of high-quality work requests. The stakeholders should be able to view the data analysis team as SMEs.

All teams within an organisation vie for budget and quality work. Creating the right impression helps to build trust with the senior stakeholders that, with the right amount of investment in the team, the business can meet its goals effectively.

Data analysts should share technical knowledge with stakeholders when requested. This may enable stakeholders to perform some tasks by themselves, create an appreciation among them of the work involved to complete tasks and raise the profile of the team. This all is possible by:

SUMMARY

In this chapter, we reviewed the key industries where data analysts work, the affiliated roles and the differences between the data analyst role and other roles. We also looked at where the data analyst could add value during a typical product life cycle and the nature of solving business problems. Data is a key influencer in the data analyst role, and we explored the effect of varying degrees of data availability on analysis deliverables. We also discussed the three types of skills required – functional, technical and soft – and looked at soft skills in particular.

In the next chapter, we will discuss the tools, methods and techniques that data analysts need to use.

1 The Basel committee was established by 10 leading countries in 1974 to deal with problems in the financial world. It publishes norms that most central banks follow to manage their credit policy.

2 More information can be found at www.sfia-online.org/en

3 The term for a class of computer bugs, relating to the storage of calendar dates for the dates beginning in the year 2000, that threatened to stop computers functioning normally.

4 This segregates customers into promoters, passive supporters and detractors, based on customer experience, and helps to predict business growth.

5 This is a term used by a section of industry for a process that captures customers’ expectations and preferences.