Data driven decisions – summary so far

At DataQuarks, we are truly passionate about data and analytics and have written about various aspects of data driven decisions. Recently a reader suggested writing a super blog post to combine various topics we have discussed until now. Thanks to that excellent suggestion, we are writing this blogpost to summarise our views on all the topics we have written in this quarter. We can definitely see this becoming a recurring exercise. It not only helps our visitors get a high level picture of our direction, it also helps us recap our work to shape our product backlog for the forthcoming quarter.

Data Utopia – steps to reality

In this blogpost, we looked at how to move an organisation from getting plenty of reports about past performance to use data and insight to decide every strategic step in the organisation. Let the business needs and not IT buzz words such as Big Data and Data Lakes drive the change. For starters, you need a framework to identify what the potential options in each decisions are and score these options based on perceived outcomes. You are better off building the models in natural business language to avoid to and fro between decision makers and analysts, which leads to inefficiencies in the decision making process. Setting up a workflow using these models into the decision making process means the buy-in both from senior and middle management.

Embrace the dark data matter

As discussed here, it is an universal problem that despite lots of investments in BI and DWH, business users prefer Excel spreadsheets and end up building plenty of spreadmarts and shadow BI/data systems. This affinity stems from the need for flexibility and to prevent too much dependency on IT. An analytics sandbox is extremely useful in this case, where IT can control the core corporate data going into the sandbox. This also enables business users to plugin other relevant data and build business rules using business language. This also eliminates the need for spending too much on sprint planning and technical development time. While there is desirable flexibility for the business users, IT has got a lot to gain as well given the governance and prevention of data leakage problems.

5 key reasons why your organisation does not need collaboration in decision modelling

While we want every organisation in this world to move towards collaboration decision modelling using data, there are many instances where this may not be needed as discussed in this blog.

  1. In organisation structure where competition between departmants are encouraged or where for compliance reasons employees cannot collaborate with each other
  2. At companies where IT and business users are aligned 100% in their vision and capabilities
  3. At organisations who are working on unique business problems each time and hence there is no need to reuse a model
  4. If an organisation has a large team of spreadsheet experts who don’t want to move to a different platform
  5. If the operational source systems have enought insights and intelligence capabilities built in

On a level playing data field for a small business

In this blogpost, we shifted our focus slightly from generic data analytics to how it could help small businesses particularly. Small business owners are always up against the big corporates in many aspects of business operations. But even small businesses can leverage the power of innovation to start the data driven journey. It is important to not postpone and start with some meaningful data available within the organisation. Having said that, start one and only from a business perspective and don’t get too excited about the technology. Many vendors including us offer free trial periods, leverage that and don’t spend until you absolutely have to. last but not the least, if you are a small business, it is better off going down the route of self-service modelling than expensive route of using technologies that depend on IT experts.

Use cases for What-If scenario modelling

We have recently kicked off a five part series to explain the powerful use cases of What-If scenario modelling for decision making. It is commonly wrongly associated only with macro economic scenario modelling, but has many uses in the micro aspects as well. Despite the hundreds of use cases we are writing about the following five.

  1. Financial Analysis
  2. Pricing
  3. Logistics planning
  4. Evidence for business proposals
  5. Customer modelling

Financial analysis is the most important use case given that finance department loves numbers and spreadsheets. It is a growing trend in not only large enterprises, but also smaller companies and start-ups. What-If scenarios help to quantify the impact of external factors. In most cases there are plenty of internal impact factors, hence multi variate models where the first step is to start with a guestimate for variables and build a probability distribution. While results can’t be exactly predicted, the process of planning various scenarios and identifying mitigating options will help prevent a shock.

The price at which a company sells its products and services has a huge impact on profitability. The second part of the series focused on pricing. The ability to mash-up multiple data sources for competitor prices, macro economic trends and internal customer surveys is a very important feature while building What-If models in this area. Pricing Analysts also need the flexibility to experiment with various input factors to understand the impact of buying behaviour.

Image Source –  by Daniel Lobo on Flickr