BASiCSNOW: Value of data is only as good as the analytical framework

October 29, 2020by Anitha Smith

Data is generated and recorded through every digital action. To understand the scale, Statista reports that data has grown from 2 zettabytes (2xbillion terabytes) in 2010 to 59 zettabytes in 2020, and is projected to reach 149 zettabytes in 2024. But data generation is only half the story. To stitch together a coherent storyline, there needs to be an analytical framework. If you are looking for a great source of solutions for real-world challenges in data analysis, we strongly recommend following Avinash Kaushik and subscribing to his blog Occam’s Razor.  His ‘Impact Matrix’ is an excellent framework to follow.

Here we discuss the 6-steps to develop an analytical framework.

  1. Set the context: Context guides the conclusions. It also filters out the irrelevant data points from the relevant ones. If the context is to understand the cost effectiveness of different marketing channels, revenue generated by each of them is a relevant data point. But how long each of these channels has been in deployment – while a valid data point – is not relevant to understand the cost effectiveness. A recently deployed channel can be far more cost effective than a channel that has been in use for many years. We recently completed an exercise for a company with a business-to-business trade model. Their traditional business development channels included trade shows, Yellow Pages, and few advertisements in newspaper supplements and trade directories. While it is difficult to fully quantify the returns from these channels, it was evident that Search Engine Marketing is more cost effective than any of these channels to bring in qualified leads.
  1. Define the objective: It is important to set the business objective right at the start of developing the analytical framework. It differs from context. To achieve a long-term objective, the analysis can be done with multiple short-term contexts. To achieve the long-term objective of being “#1 profitable brand in a category”, cost effectiveness of marketing channels is one context, prioritisation of customer segments is another context, optimised supply chain is a 3rd These 3 contexts need to be framed within the objective of “#1 profitable brand”. S.M.A.R.T. is a useful framework to define objectives. One of our clients stated his objective as “to be the 1st choice for quality products”. ‘Quality’ being not a numerical metric alone; we developed a framework that included both quantitative and qualitative metrics.
  1. Choose the variables: Digital ecosystem is inundated with data points. It is therefore easy to get lost while developing an analytical framework. The context guides the variables so that the right conclusion can be arrived at. Differentiate between metrices, KPIs (Key Performance Indicators) and KSOs (Key Success Outcomes). All KPIs and KSOs should be set against targets, achievable within the timeframe. In the context of cost effectiveness of marketing channels, KSO can be achieving +x% ROI over the previous year. The KPI is +x% revenue generated. Both targets should be based on historical data. The metric to be measured at a set frequency (daily/weekly) in this case is CPM or CPA.
  1. Plotting the framework: Plotting a framework is different from plotting a graph. Graph is bound by axes while framework has matrices, like the 2×2 matrix of Impact Matrix. Matrix provides flexibility to bring together variables with different dimensions. You can have a framework with %s, absolute numbers, ratings, gradings, and qualitative descriptors. You can also use numerical equivalents of qualitative descriptors. Net Promoter Score is a good example of a qualitative variable reflected numerically. When we crated the framework for our client focused on ‘quality’, we chose variables such NPS but also included purely qualitative variable like a grading for taste.
  1. Statistical analysis: It cannot be stressed enough that to develop a good framework, enough rigor has to be applied to the analysis. Only then the nuances are revealed. Top level analysis usually provides answers that are already known or hypothesised. But ‘digging deeper’ through advanced statistical analysis, more insights can be generated to unearth the ‘unknown unknowns’ (another term used often by Avinash Kaushik). There might not always be an internal resource with advanced data skills set. You can onboard a freelance data analyst, but there are concerns of confidentiality. It might be worthwhile to train one of the internal resources who is comfortable with data. There are many free online courses that teach 1st level advanced statistical analysis.
  1. Visualisation: This again cannot be stressed enough. It is not about making ‘good looking’ charts. It is the art of telling a story through numbers. Like any story, you need to first decide on the target audience. Visualisation for a Finance team should be different from the one used for Senior Management. Free tools like Google Data Studio has made this task easier. There are also free Infographic generators. Spend time to craft the story and then use tools to tell the story visually through data.

An important aspect digital transformation is the data infrastructure. At PHYGiTALNOW, we built our strategic framework based on a data. Through C.A.R.T. we analyse the business through 4 contexts – Customer, Architecture, Reach and Transact – to achieve the long-term objective of ‘sustainable growth’. But we recognise that different industries measure their ‘growth’ differently. We therefore develop bespoke frameworks for each industry. Please reach out to us at and we can help build an analytical framework that is specific to your long term objective.


PHYGiTALNOW is a digital marketing consultancy in Dubai focused on transforming businesses for sustainable growth. The primary focus is to realise the full revenue potential by providing best-in-class Digital Marketing Services – Search Engine Optimisation, Search Engine Marketing, Website and App Development, Social Media Strategy and Activation, Email Marketing and Digital Advertising. We also provide E-commerce specific services – Online Marketplace Management, Conversion Optimisation and Basket Analysis. We tailor-make our consultancy packages depending on the digital maturity of the business.