RESOURCES / KNOWLEDGE CENTER
Analytics FAQs
What is Predictive Analytics?
What are the Signs that my Business can benefit from predictive analytics?
What can predictive analytics help me do?
What kind of data would you need to do an Analytics Exercise?
I do not have a large datawarehouse – will Analytics work for me?
Is there a quality expected in the data available with us?
Do you deliver dashboards as your final product?
Will we need to know advanced statistics / modeling techniques to make sense / use of the model outputs?
Would we be able to tangibly measure the improvement from an Analytical model?
How does analytics help in making better decision that a domain expert cannot already make?
What are the insights that you can present that our business experts would not know already?
How long would it take me to train my resources to develop internal capability to use the models?
How does LatentView typically engage with a client?
What are the benefits of the Global delivery model?
How can LatentView add value if I already have an in-house Analytics team?
What are the different domains that LatentView has worked in?
What Software does your company use for executing Analytics projects?
What Analytics techniques are used by LatentView?
How do you ensure that our Data is secure with you?
What is the minimum success guarantee for an Analytics exercise?
Q. What is Predictive Analytics?
Predictive Analytics is the science of predicting future events using data mining and statistical techniques on historical data. Predictive Analytics encompasses a range of mathematical techniques to derive insight from data with one clear objective: Identifying the best action for the situation to give the best possible benefits.
Q. What are the Signs that my Business can benefit from predictive analytics?
Q. What can predictive analytics help me do?
Q. What kind of data would you need to do an Analytics Exercise?
Data only becomes useful to us when it can be moulded into actionable information. We typically use data across five broad dimensions
Q. I do not have a large datawarehouse – will Analytics work for me?
Analytics can work on smaller datasets (even as few as thousands of records). The number of records for statistical significance depends largely on the Objective in hand and the kind of analysis that needs to be carried out. We have carried out analytics exercises with around 5000 observations and delivered satisfactory results.
Q. Is there a quality expected in the data available with us?
Data quality is directly proportional to its utility in predictive analysis. However there are methodologies that can be used to remove inconsistencies in data, clean the data for analytical modeling. As part of the Analytics exercise, we would partner with the client to identify how we could better capture, maintain and use the data for future modeling / analysis.
Q. Do you deliver dashboards as your final product?
Dashboards are a part of our solution but not the core of it. Our core solution is in mining data to provide the necessary insights and foresights to business users to make the right decisions. Dashboards could be custom built to illustrate and track the benefits of the analytical solution.
Q. Will we need to know advanced statistics / modeling techniques to make sense / use of the model outputs?
No. The results derived from our analyses are presented in forms of easily interpretable charts and tables that require no expertise in statistics/modeling to comprehend. Further, the inferences from the charts/tables are worded out for the Business users.
Q. Would we be able to tangibly measure the improvement from an Analytical model?
The performance of the model is measured by a test and control group approach. The model is used to take decisions on the Test group whereas the current business rules are used in the Control group. The difference in performance in the Test and control group for a decided performance window is taken and attributed to the model performance.
Q. How does analytics help in making better decision that a domain expert cannot already make?
Analytics performs the role of a ruler in the hands of a design architect. While an experienced architect can draft layouts at ease without a ruler, a ruler can take the layout to a different level of accuracy.
Q. What are the insights that you can present that our business experts would not know already?
Business experts are usually very good at analyzing one or two dimensions at a time. For example, customers in the age group 25 – 30 would be the target for a cell phone with the latest technology.However when the number of dimensions is increased, it would become increasingly difficult to identify and differentiate between segments: taking characteristics like income, age, occupation, locality and presence of a mortgage. This is where statistical analysis comes into play which helps distilling insights for the Business user by looking at multiple dimensions in the Data (multi-variate analysis).
Q. How long would it take me to train my resources to develop internal capability to use the models?
We would decode the statistical intricacies of the models into plain mathematical equations / business rules that can be easily integrated into your current IT system / applications. Our experience with clients in different sectors has shown that the time required to integrate the analytical model and develop the internal capabilitiy to use them is usually minimal.
Q. How does LatentView typically engage with a client?
We operate in a Global delivery model with our business analysts working closely with the business team onsite to define the business problem / modeling objective. The development of the analytical models is usually executed from our offshore development centre (Chennai).
Our engagement model is flexible to the client’s needs / demands and we can fine-tune our delivery model to match client’s expectations.
Q. What are the benefits of the Global delivery model?
Q. How can LatentView add value if I already have an in-house Analytics team?
Q. What are the different domains that LatentView has worked in?
LatentView has a wide variety of experience across different sectors. An illustrative snapshot of our experience is listed below:
Q. What Software does your company use for executing Analytics projects?
Typically we use SAS for a majority of our analytics / modeling requirements and RDBMS software for data management. However, we use specialized software like Bonmin, Libsvm etc. for high end modeling requirements. We also actively evaluate and use Open source software like R, Weka depending on the client’s needs, software’s flexibility / constraints, ease of integrability and deliverable formats.
Q. What Analytics techniques are used by LatentView?
LatentView has expertise across range of Analytics techniques around Classification, Clustering, Association, Forecasting, Descriptive Analysis and Optimization. For more details refer to
Q. How do you ensure that our Data is secure with you?
LatentView has a secured environment with different levels of access given to different users to ensure data is secure within LatentView premises. Additionally, all personally identifiable information is either masked / encrypted to ensure that no customer information can be leaked. LatentView has a Data Security and Privacy policy and can be made available to a client / prospect on demand.
Q. What is the minimum success guarantee for an Analytics exercise?
LatentView, in its experience, has been able to deliver significant benefits to the client through implementation of its analytics solutions across various functions. The success of an Analytics exercise depends on the number of dimensions, quality and availability of data with the client. However even sparsely populated data across few dimensions can be used to develop models that give better insights into the business than done by any non-analytical approach.
Analytics FAQs
What is Predictive Analytics?
What are the Signs that my Business can benefit from predictive analytics?
What can predictive analytics help me do?
What kind of data would you need to do an Analytics Exercise?
I do not have a large datawarehouse – will Analytics work for me?
Is there a quality expected in the data available with us?
Do you deliver dashboards as your final product?
Will we need to know advanced statistics / modeling techniques to make sense / use of the model outputs?
Would we be able to tangibly measure the improvement from an Analytical model?
How does analytics help in making better decision that a domain expert cannot already make?
What are the insights that you can present that our business experts would not know already?
How long would it take me to train my resources to develop internal capability to use the models?
How does LatentView typically engage with a client?
What are the benefits of the Global delivery model?
How can LatentView add value if I already have an in-house Analytics team?
What are the different domains that LatentView has worked in?
What Software does your company use for executing Analytics projects?
What Analytics techniques are used by LatentView?
How do you ensure that our Data is secure with you?
What is the minimum success guarantee for an Analytics exercise?
Predictive Analytics is the science of predicting future events using data mining and statistical techniques on historical data. Predictive Analytics encompasses a range of mathematical techniques to derive insight from data with one clear objective: Identifying the best action for the situation to give the best possible benefits.
Quick decisions need to be taken in a dynamic market environment, weighing a lot of factors
You know more about what your customers have done than about what they will do
You want a consistent and optimized decision process that helps you differentiate customer treatment. There’s a risk associated with poor decisions—and a reward for making better ones
Know your customers—and act on the insight
Work smarter and faster by replacing guesswork with science
Reduce costs with a clear gauge of risk and uncertainty
Add consistency to business decisions, improving compliance and customer service
Q. What kind of data would you need to do an Analytics Exercise?
Data only becomes useful to us when it can be moulded into actionable information. We typically use data across five broad dimensions
Transactional Data – Purchase, Payments etc.
Demographic Data – Age, Income, Occupation etc.
Interaction Data – Customer Moments of truth through different channels
Attitudinal Data – Survey Data, Social Media etc.
External Data – Macro-economic indicators, zip code demographics
Analytics can work on smaller datasets (even as few as thousands of records). The number of records for statistical significance depends largely on the Objective in hand and the kind of analysis that needs to be carried out. We have carried out analytics exercises with around 5000 observations and delivered satisfactory results.
Data quality is directly proportional to its utility in predictive analysis. However there are methodologies that can be used to remove inconsistencies in data, clean the data for analytical modeling. As part of the Analytics exercise, we would partner with the client to identify how we could better capture, maintain and use the data for future modeling / analysis.
Dashboards are a part of our solution but not the core of it. Our core solution is in mining data to provide the necessary insights and foresights to business users to make the right decisions. Dashboards could be custom built to illustrate and track the benefits of the analytical solution.
No. The results derived from our analyses are presented in forms of easily interpretable charts and tables that require no expertise in statistics/modeling to comprehend. Further, the inferences from the charts/tables are worded out for the Business users.
The performance of the model is measured by a test and control group approach. The model is used to take decisions on the Test group whereas the current business rules are used in the Control group. The difference in performance in the Test and control group for a decided performance window is taken and attributed to the model performance.
Analytics performs the role of a ruler in the hands of a design architect. While an experienced architect can draft layouts at ease without a ruler, a ruler can take the layout to a different level of accuracy.
Business experts are usually very good at analyzing one or two dimensions at a time. For example, customers in the age group 25 – 30 would be the target for a cell phone with the latest technology.However when the number of dimensions is increased, it would become increasingly difficult to identify and differentiate between segments: taking characteristics like income, age, occupation, locality and presence of a mortgage. This is where statistical analysis comes into play which helps distilling insights for the Business user by looking at multiple dimensions in the Data (multi-variate analysis).
We would decode the statistical intricacies of the models into plain mathematical equations / business rules that can be easily integrated into your current IT system / applications. Our experience with clients in different sectors has shown that the time required to integrate the analytical model and develop the internal capabilitiy to use them is usually minimal.
We operate in a Global delivery model with our business analysts working closely with the business team onsite to define the business problem / modeling objective. The development of the analytical models is usually executed from our offshore development centre (Chennai).
Our engagement model is flexible to the client’s needs / demands and we can fine-tune our delivery model to match client’s expectations.
Helps clients manage their talent pool to effectively meet their Analytics delivery needs
Balancing Work Overflow: Typically clients have spikes in requirements for analytics services and LatentView offshore team would help the client to better manage these spikes
Do more with Less: We provide high value analytics services through competitive and flexible pricing models
Specialized Analytics provider: Partnering with a specialized provider like LatentView Analytics would bring in cross-domain and cross-functional exposure which is key to high quality deliverables
Focus: Client can outsource non- core but important work like Model Maintenance, fine-tuning, generating ad-hoc reports etc and use their resources to focus on core initiatives
LatentView has a wide variety of experience across different sectors. An illustrative snapshot of our experience is listed below:
Insurance - Campaign Management, Compliance Risk Analytics, Lapsation Analytics, Repeat Purchase Analysis, Agent Retention
Business Process Outsourcing – Collection Analytics, Real-Time Recommendation Analytics
Financial Services – Market Segmentation, Activity Scoring
CPG & Retail – Campaign Management, Media Mix Modeling, Promotions Modeling
Typically we use SAS for a majority of our analytics / modeling requirements and RDBMS software for data management. However, we use specialized software like Bonmin, Libsvm etc. for high end modeling requirements. We also actively evaluate and use Open source software like R, Weka depending on the client’s needs, software’s flexibility / constraints, ease of integrability and deliverable formats.
LatentView has expertise across range of Analytics techniques around Classification, Clustering, Association, Forecasting, Descriptive Analysis and Optimization. For more details refer to
LatentView has a secured environment with different levels of access given to different users to ensure data is secure within LatentView premises. Additionally, all personally identifiable information is either masked / encrypted to ensure that no customer information can be leaked. LatentView has a Data Security and Privacy policy and can be made available to a client / prospect on demand.
LatentView, in its experience, has been able to deliver significant benefits to the client through implementation of its analytics solutions across various functions. The success of an Analytics exercise depends on the number of dimensions, quality and availability of data with the client. However even sparsely populated data across few dimensions can be used to develop models that give better insights into the business than done by any non-analytical approach.
RESOURCES
CONTACT US
