It’s no accident that the more successful ‘new business’ clients I have worked with have been those that have used analytics to gain a good appreciation of what’s happening over time. Hence, it’s surprising that analytics is something that’s often forgotten by people developing their new product or service.
Analytics are sadly nearly always the lowest priority during development but this suddenly changes towards the end of a project or after release when the product owner realises their oversight because they can’t see what’s going on. Analytics is something that is always more difficult and more costly to add afterwards.
This article explores types of analytics, gives some pragmatic data collection tips and takes a deeper look at some considerations when collecting data.
Why Use Analytics?
Analytics tells you where things are succeeding and failing. Knowing where, leads to why which allows you to maximise what’s working and improve or remove what’s failing. Spot and act on both positive and negative trends. Fine tune what to concentrate on based on fact rather than on some manager’s incorrect assumptions.
Data can also be used for proof, for example, proof of effectiveness of a campaign or regulatory aspects such as proof of compliance. It can also provide much needed proof of worth and ROI to leverage subsequent investment and funding.
Data is very powerful. Output data and advanced learnings from analytics can sometimes become an asset in itself that can be re-sold and used by others.
Types of Analytics
There’s lots of hype over IoT, big data, data science, machine learning but in the end they are all about collecting the data and doing some analysis. Data tends to be usage data and error data. Usage data also includes performance data. Here are some types of data that can be collected:
Traditionally, most analytics data has been metadata i.e. data about data. However the physical category is becoming more prominent and useful. This is the focus of my ‘new business’. Knowing the presence, location and physical quantities such as open/closed, temperature, humidity, movement, light level etc. allow for deeper and more useful insights that can directly impact organisations’ efficiency.
Apart from perhaps web and app analytics, don’t collect data for the sake of it. Instead collect data where there’s most value. Work out the key data points that, if collected, will achieve the greatest insight or value.
What kind of analytics provide the most value? Think about data describing bottlenecks, human effort-limited tasks, costly workrounds, stoppages, downtimes, process delays, under-used equipment and under-used people. Can you measure these things and predict they are about to happen?
Don’t ignore situations that initially might seem impossible to analyse. AI machine learning is a great way to unearth insights that have been ignored for years or decades because they were thought to be too difficult or insolvable. Technology might now be able to solve these cases. If you are one of the few able to extract these insights then you have engineered a competitive advantage.
For analytics used for support purposes, it’s best to have some kind of ‘switch’ so you can turn deeper analytics on and off.
- Monitor – Many companies put analytics in place but rarely look at the output. Continuously monitor the output. Use alerts to notify unusual situations.
- Protect Privacy – Balance the need for data against information privacy. Comply with data protection regulations and ensure systems don’t leak information? If you use 3rd party tools and platforms ensure they don’t have privacy issues.
- Review Functionality – System and processes change over time. Ensure the analytics continues to do what’s needed.
- Protect Data – Ensure the securing of APIs. Regularly patch systems to help protect against vulnerabilities.
- Monitor Cost – Storing and transferring data usually has a cost. Monitor the cost over time so that it doesn’t cause bill shock.
- Ensure Continuance – Using 3rd party tools and platforms causes dependencies that you have less control over. Put in place mitigations should a 3rd party service change, fail or go out of business.
While I have been looking into the practical and pragmatic aspects of business analytics, it’s a subject in its own right with degree courses and books. There also also processes that can be followed.
Design-in appropriate analytics to continuously monitor your new business. Detect problems quickly and identify processes that work well and those that don’t. Replace guesswork and gut feeling with quantitative measures. Analytics is the key ingredient of providing for efficiency and hence improve business performance.