25 Tips for Managing Text Analytics Projects

 

Sometimes text analytics project in your company may not go the way you envisioned for multiple reasons. It could be bad data, incorrect software, stakeholder issues etc.

You may ask how to make text analytics projects run correctly for your business?

The following are general phases in a text analytics program with highlighted tips for planning and overall delivery.

1.2 Process

 Defining Business Objectives

  1. Clearly state business objective
  2. Focus on particular business questions like most profitable customers
  3. Ensure stakeholder awareness
  4. Choose the right KPIs
  5. Identify needs and deliver to appropriate personnel.

Creating Analytics Brief

  1. Document all project activities
  2. Check for data supporting your analysis
  3. Create a project plan, agree the timescales of the project, allow time to review the results and build in a contingency for anomalies (typically 20%)
  4. If data availability is an issue, inform stakeholders of revised approach.

Use the Right Resources

  1. What is the budget?
  2. Verify data is legally compliant for use in analysis
  3. If the data needs to be sent to a third party are the relevant data processing agreements in place? Do they have appropriate data security standards/appropriate legal contracts?

Look for Good Data

  1. All data should be verified
  2. Run basic cross checks of data quality with respective department
  3. Validate counts with existing reports
  4. In-depth analysis starts when data is signed off

Never Deviate From Main Objective

  1. Ensure analysis relates to the brief
  2. Get regular updates from analysts
  3. Doubtful findings need review before inclusion in final report

Generating Valuable Deliverables

  1. Ensure factual presentation of the results
  2. Allow questions
  3. Stakeholders agreement on the most business critical findings
  4. Agree on prioritization of implementation plans and initial timescales
  5. Identify additional analysis or market research is required
  6. Ensure anything that you learn in the process about planning, techniques, or findings are shared with other analysts for future projects.

 

Effective data visualization and dashboards

Data Visualizations: Making the Value, Happen!

Effective data visualization tells the user, at a glance, everything that is necessary to know. Hence the dashboards need to be well designed so as to give the users the right picture to analyze data, track performance and make informed decisions. Therefore, it is important to know the dashboard best practices so that users have access to what they really need.

Designing an effective dashboard

Planning is essential for effective dashboards. It is important to involve the business users when mapping the delivery requirements. It is a good idea to host a planning interview with the business users where the key business metrics that need to be tracked are identified. In addition, the business users should share the frequency of information updating. The data visualization team should be aware even before the start of the project about whether the metrics definitions are the same across the business unit and also whether the data level access rights etc. All of this will help in the creation of dashboards that are both user friendly and deliver the value promised.

It is important that the users are not overwhelmed with information. Depending on the user roles, the information access should be streamlined. It is most often enough to provide high level metrics that highlight the business process performance. In most cases, the business users require summarized information on the dashboard screen and this should be visualized in such a manner that it is easy to understand at a glance. If interested, the users should be able to drill down further information on the metrics shown.

Best practices for data visualization:

  1. There should be around 4 to 6 metrics on a single dashboard page. The metrics pertaining to a specific business unit need to be on a specific tab or page. This helps provide easy access and understanding.
  2. The ability to drill down should be a maximum of up to three levels. If the user needs to drill down further, they can always take the assistance of a data analyst or the IT team when needed.
  3. Charts, tables, graphs etc. should be utilized for data visualization. This helps in easy interpretation of the data. It is better to not go for pie charts. Bar charts are a better option to showcase differences to the users. The usage of gauges also needs to be restricted, as they do not provide too much information and at the same time occupy a lot of screen space.
  4. Dashboards need to be interactive; this would help the users access information easily and also customize the data views. Dashboards with filters help users to maneuver the data scope so as to meet their specific needs. Alerts help in capturing the user’s attention.

Data visualization mistakes to avoid

It is important that the users are not overwhelmed with information. Depending on the user roles, the information access should be streamlined. It is most often enough to provide high level metrics that highlight the business process performance. In most cases, the business users require summarized information on the dashboard screen and this should be visualized in such a manner that it is easy to understand at a glance. If interested, the users should be able to drill down further information on the metrics shown.
The graphs and charts should not be poorly designed. Considerable thought should be put into creating graphs and charts which boost viewer comprehension. Garish colors should be avoided.
Using metrics that the broader audience would not understand should be avoided. Easier concepts and metrics would help in keeping the dashboard reader friendly.

Keeping these simple tips in mind while working on data visualization and dashboard design would help in coming up with solutions that are coveted by users and help in business decision making.

Who Handles Customer Analytics

Who handles your Customer Analytics?

Who Handles Customer Analytics

Processing all the customer data and creating actionable customer analytics is a very complex task. It requires coordinated efforts by skilled and specialized professionals working across multiple departments of a company; particularly customer experience, IT and marketing. However, the division of task varies from company to company depending on the scale of customer intelligence operations present in the organization.

In order to get to the bottom of who should ideally manage customer intelligence, we need to first understand what steps customer intelligence operations essentially encompass. It includes –

  1. The storage wherein the huge chunk of customer information extracted from various social media platforms, customer feedback, and other sources are stored together. This information can be structured or unstructured.
  2. The next step is processing the customer data to segregate the relevant from the irrelevant.
  3. The third step is analyzing. Once the processing is done, the relevant information is scrutinized for actionable insights to formulate further business strategies.

Among these activities, storage and processing go hand in hand. The servers capable of processing the customer data require technical expertise to handle it. Hence, many customer intelligence professionals work in the IT department. Some companies may have their analysts working in business functions that use customer data to gain insights and make decisions. Since customer intelligence is still very much in its evolution stage, many companies may not actually have a separate group that consists of a team of professionals designated to solely overseeing its processing and analyzing.

Let us see some of the typical customer intelligence organization –

Customer analytics experts in business functions like marketing –

Customer intelligence experts working in marketing departments have an upper hand when it comes to modeling plans and policies more in sync with customer preferences. This is because the key to attracting customers and building customer loyalty lies with them. The customer data can be analyzed for insights that can be used to achieve more personalization in customer services. Marketing campaigns with a sharper focus on customer needs designed through analytical insights would reap far better results than those designed vaguely, without the highly reliable analytical data support to rely on. Since marketing functions revolve around the business, they generally don’t have customer intelligence experts handling the processing. That majorly remains the prerogative of the IT department.

Customer analytics experts in IT functions –

Since customer intelligence is very much a technology-driven process, in many companies, the IT department is the one where the core function is located. Technology is the backbone of customer intelligence. Customer intelligence professionals working in the IT functions know exactly the kind of technological infrastructure that is required to process customer data. The analysts can make recommendations for high-tech upgrade, weighing the scalability and the flexibility of the customer data with them. The needs and expectations of the analysts and the technology must match at all levels, so as to ensure that efficiency in processing is always maintained. They have to be mindful of the costs, reliability, packaging and ergonomics along with the pre-requisite technical knowledge.

Customer intelligence professionals working in different functions, having an expertise in their distinct capabilities, need to maintain a coordinated working pattern and a lucid communication channel to process and analyze the customer data for actionable insights.

Customer analytics experts in a separate group-

Some companies understand that understanding of the customer, creating actionable insights and utilizing them towards business goals is an interdisciplinary activity. A separate group of experts who understand data, technology as well business are most suited not only to bring together all the customer data, but also to transmit the customer intelligence to appropriate teams within an organization. This leads to better results and more success. Companies with a separate function generally had a higher ROI than those who did not.

Centralization of customer intelligence operations helps in multiple ways-

1) The experts can offer unbiased opinions strictly on the basis of customer intelligence, without intervention of vested interests of other company functions; whose judgment might be clouded in favor of its department’s well-being or constrictive due to a department-specific limited scope of understanding.

2) The customer intelligence professionals, together with their freedom, get a sense of assurance of job security and scope for progress, which is otherwise limited to other business functions due to the hierarchical structure of the businesses.

3) It helps in providing a singular view of customers, which is of absolute necessity for delivering better customer experiences. It saves the confusion that often arises with disorganized and multiple versions of customer information present in different departments.

4) It squashes the chance of being limited by traditional thinking. This is because it gives much space and independence for innovation and technological evolution outside the realm of typical business bounds.

5) It builds high levels of trust between the data scientists, who present insights on customer data, and the functional managers.

Having a separate centralized group of customer analytics professionals along with analysts in specific business units would establish a more cohesive, stable and healthier system of analytics communication. The analysts should be able to put in place the recommendations put forth by the central group, whose main task would be processing rigorous data and arriving at business intelligence through its data science capabilities.

What is your take on this?

Avoid These Big Data Mistakes

Big Data Mistakes to Avoid

Avoid These Big Data Mistakes

For a company, initiating into big data is a huge step, something that has the potential to transform the face of any business in the market for the better. With the ongoing expansion and adoption of analytics by a majority of business houses of various industries and the availability of an even larger amount of literature on big data in the virtual world, it is easy to get confused or lost by erroneous perception created due to a careless assessment.

Every business requires a different business module which works specifically and uniquely for that particular company. Processing big data requires unique strategic analysis that varies from one company to other depending on its market status, customer demographics and extent of social media presence, among others. Getting gripped by misguided notions can lead to unfulfilling big data outputs causing great disappointment in the big data project. However, by avoiding 5 typical mistakes in the process you can ensure the actionable insights and results that big data ambitiously promises.

Selecting wrong KPIs

Key performance indicators are aplenty in the social media realm. With big data, you have at your disposal an array of KPIs like sales growth, product performance, customer attrition, inventory, turnover etc. to scrutinize. Due to this, there are chances that a company might go wrong in focusing on the issue-specific KPI and get incorrect results. Here, the question of exactly what to measure is important to identify the kind of KPIs to use and analyze. It is therefore dependent on the objectives of the business plan and the expected outcome.
Two mistakes can lead to incorrect analysis:

  1. Having all the data in hand but asking incorrect questions.
  2. Asking the correct questions but not looking into appropriate data for answers.

To avoid these, there is a simple 3-step process that you can follow:

    1. Formulate the right questions by keeping the business objectives in mind;
    2. Sift through all the data and extract relevant bits that will give actionable insights; and finally
    3. Analyze them to model strategic solutions.

No effort to unify dispersed data

The stagnant data silos can impede the data management and data integrity efforts of an organization. It can hamper productivity by coming in the way of organized data utilization and processing. The repository data is mostly unused, lying in some IT nooks that do not participate in the integrated voice of customer analysis. Ignoring internal data, which exists in several legacy systems and other internal data sources like surveys, call center notes, CRM etc. is a huge misstep.
For a complete data integration, it is important to build a 360-degree Voice of Customer Intelligence. It enables the integration of both the structured and the unstructured data which gives a complete view of the ecosystem under study.

For effective intelligence, this integration is needed.

Lack of inter-departmental synergy

The ‘Silo Mentality’ as defined by the Business Dictionary is a mindset present when certain departments or sectors do not wish to share information with others in the same company. This type of mentality will reduce efficiency in the overall operation, reduce morale, and may contribute to the demise of a productive company culture.

Any organization inevitably requires a coordinated working of different departments to reap results, due to which the presence of organizational silos proves to be a highly detrimental barrier to deal with. It’s a major roadblock for a data-driven business.

Tackling this, however, is easier than believed. The CMO-CIO synergy needs to be strengthened as the first major step towards breaking organizational silos. More often than not, what we see is a huge gap between the IT and marketing department, causing the lack of understanding and disorganized, haphazard flow of information between the two. Maximum efficiency can only be ensured by bridging the gap between them.
Other simple steps like:

  1. Bringing on board a Chief Data Officer,
  2. Enhancing Marketing-Technology synergy,
  3. Focusing on critical/relevant data etc. can help in breaking the organizational silos and maximizing a company’s productivity.

Opting for free/cheap big data tools

It’s true enough that the market today is full of free software providing easily accessible analytics to companies and promising complete integrated solutions to business problems. While not debunking their utility entirely, it is very important to understand that their functions serve only a limited purpose and in no way are enough to provide a wholesome analytics solution for the company’s specific requirements. Falling for the ‘free service’ tag can cost the organization its security and even present a possible threat of losing some of their precious data. This is because most open source software aren’t fully secured and are not known to scale well to the mammoth data analytics requirements.

A company, therefore, must firstly understand and define its objectives and requirements and then accordingly go for a solution provider who offers them the customized services.

Letting the big data solution drive the results instead of the business objective

All the big data and analytics center on business objectives at the end of the day. The extraction, processing and analyzing are all done with the intention of solving business issues which invariably lie at the core of all the ventures of the company, including the most technologically demanding big data project. The CIO must streamline all his IT workflows with the marketing goals. This alignment between technology and business is of utmost importance because excess focus on only the technological front without periodic business intervention and review would produce incongruous data and hinder productivity.

Using social business intelligence services can ensure high ROI and tangible outputs. A synergy of data scientists, consultants, processes and technologists is a must for an efficient working and delivery of customized, problem-specific solutions to business issues.

Are You Facing These Big Data Challenges?

Big Data Challenges

While leading companies are banking on big data technologies to tackle the changed reality of the digital age, others find the whole big data bandwagon a little overwhelming. It is no secret that companies generate huge amounts of data, and insights created from this data can lead to several benefits, such as a better understanding of customers, better customer experience, improved marketing effectiveness, higher sales driven by a better understanding of your prospects, and more effective market research.

However, such companies have several concerns before going for the implementation of a big data program. These might dampen a company’s interest in investing in big data.

The following are some of the common concerns companies have when it comes to investing in big data.

Cost of big data program

A big data program is a technology-driven process. Implementing a full-fledged program may incur heavy expenditure. There are three main cost-related factors:

  • Cost of technology: This involves the cost of advanced analytics technology which is required to aggregate and process big data.
  • Cost of hiring big data experts: This involves investments needed to build a team of big data analysts.
  • Ensuring return on investment (ROI): Most big data programs fail when the insights generated do not translate into business actions. So, how can companies ensure that the big data program pays off in terms of business outcomes?

Lack of talent with big data expertise

A big data program requires a team, which understands not only the data but also the technology along with the business. Acquisition and training of such talent involves expenditure in terms of money as well as time.

Poor quality of data

The quality of insights generated from a data set directly depends upon the quality of the data itself. Big data would render itself useful only when it does away with all the clutter that inevitably comes with it. Stale, irrelevant and garbage data leads to a huge wastage of money, time and other resources by churning out irrelevant and false intelligence.

Ethical implications of using big data

Big data extensively gathers and reviews customers’ information. Therefore, a company may have several concerns about the impact of regulations related to customer privacy.

How to address the big data challenges?

These issues might seem daunting at first. However, they can easily be dealt with. Here’s what you can do to tackle the above obstacles:

  • To get around the budgetary restriction, companies can opt for modular solutions based on their business objectives. For example, enlisting the help of an analytics company to generate a specific report like a Popular Content Report, which is delivered to you periodically based on relevant social media conversations. The insights from this report can boost marketing by establishing the business owner as a thought leader or keeping the sales team updated about the crucial information they need to know. This approach is very cost effective and addresses the business’s priorities. For more information, check out On Demand Analysis service.
  • Every company wants high ROI. VOZIQ, through its synergy of platforms, industry frameworks and experts, operationalizes the business intelligence for role-based, action-oriented alerts. This ensures that the insights generated do not sit idle but drive actual business actions from your teams. These insights-driven actions ensure that you can maximize the return on investment from the big data program.
  • Several companies, including VOZIQ, offer analytics services which include not only technology but also a team of experts which brings with them proven processes and domain experience. Such a team works as an extension of your internal teams to design and implement a custom big data analysis program. This way, you don’t have to bear the cost of expensive big data technology development and instead get the benefits from the expert solutions of the service provider.Check VOZIQ’s professional analytics services
  • The concern over data quality can be addressed through better technology as well as human intervention. For example, VOZIQ employs five advanced analytics engines which are designed to find relevance in the large pool of data. These find the most relevant business insights and KPIs about your customers and competitors. Secondly, the human intervention by VOZIQ analysts (check VOC3 Intervention Service) ensures that the quality of the data as well as the insights is of the highest accuracy and relevance.
  • For ethical implications, one must simply remember that big data is ethically neutral as long as a customer is not deceived into sharing his/her personal data. Companies can use this data to offer better customer experience without sharing or selling the data to third parties.