Blog

Why analytics is important for business


Portrait of John Hughes
By John Hughes

04 May 2018

If you're an organisation with something to achieve, you need analytics. Find out how analytics can help you better understand the past and predict the future.

Who needs analytics, anyway?

Okay, I could list a million individuals and organisations that benefit from data analysis in some form or other. In my opinion, though, it’s worth making a clear statement about what analytics is and who it helps in order for me to stress how critical it really is – and how so many people underestimate the requirements of analytics as a discipline.

So, here goes.

Analytics is the critical review of data in order to establish patterns. Individuals and organisations that have something to achieve will benefit from the use of analytics to help predict whether or not they are on course to achieve their aims. Using analytics, they can identify how they evolve their strategy to increase the chance of achieving those aims. The patterns in the data are the stories through which strategists learn.

As to who needs analytics, if you are an individual or an organisation with something to achieve, you need analytics.

Why analytics is important

Analytics gives you the power to do two things. Firstly, you can track progress so you know where you have come from. Secondly, you can predict the future, in that you can extend the patterns of data into the future.

This is incredibly important because when you have something to achieve, it tells you when you will hit your target, or even if you are on course to hit that target at all.

There is another layer to analytics, though, and that is segmentation. Segmentation enables you to look at subsets of your data based on one dimension (or a combination of dimensions) of it and to compare these subsets together.

This is really where the critical review of data is undertaken. By establishing that different segments show different patterns, we can make more complex predictions about the future. With no segmentation, we can only extend the overall data trend forward; with segmentation, we can make alternative models of the future based on actions we can take to influence the segments.

For example, a few years ago, we discovered that one client was seeing their user base transition to mobile devices much quicker that the average across other clients. When other websites were seeing 20% of their sessions delivered on mobile devices, this client was seeing 30%, and we could see the velocity with which this segment was growing. Furthermore, we could see that this segment had a much lower conversion rate than the segments for tablets and desktops.

Segmentation was helping us to predict that in the future we would have a conversion rate problem on the site. This helped build the business case for action before the problem occurred. We could recommend suitable redevelopment action immediately. By doing this, we avoided potentially 18 months’ worth of delay before the problem would have begun to surface using non-predictive methods.

How marketing analytics is changing

Traditionally, marketing has been a very creative industry, leaning heavily on the experience and intuition of a number of great creative thinkers. Advertising worked by getting a brilliant creative approach in front of as many people as possible, knowing it would at least then reach some people who mattered.

“Half the money I spend on advertising is wasted; the trouble is I don’t know which half.”

Creative thinking is still important, but as digital has become more and more a key component of organisations’ marketing activity, they have got more and more access to data. This has enabled much better critical analysis of what marketing is really working. As such, it should no longer be commonplace for an organisation to not know to some degree what marketing channels are working.

The Cambridge Analytica scandal has highlighted to the general public just how much data advertisers now have access to. In having such complex data, advertisers can micro-target their audiences, no longer relying on the scattergun approaches of old.

It’s not just through platforms like Facebook that advertisers can micro-target, either. Demand Side Platforms (DSPs) provide a way for advertisers to track and target individual users based and a huge array of data dimensions. The DSPs use their own analytics and predictive algorithms to decide which user is best served which advertisement. This means advertisers are no longer tied to bulk buying from publishers. Instead, the algorithms can predict which users are a good fit for the strategy and product and buy inventory accordingly.

Marketing is now impossible to sever from analytics and data. Data is enshrined in everything marketing does.

What about service design analytics?

However, it isn’t just marketing that has become dependent on data analytics. Service design and delivery both benefit immensely from data.

Good products and services meet users’ needs head-on. User experience practitioners collect both qualitative and quantitative data to help organisations understand user’s needs, and to prototype service design models. As our Digital Consultant, Chelsey, attests to: “We refer to analytics data all the way through our projects to understand how users engage with web services. At the start, analytics helps us get a sense of users’ top tasks and target our user research.

"It doesn’t stop there though. We also use analytics data during testing periods, such as a beta service release. Quantitative analytics data is a key piece of evidence we consider alongside qualitative feedback (from user research and testing) as we build and improve services.”

Furthermore, once services are launched, data is collected and analysed in order to inform continuous improvement, for example, by identifying where in a process users get stuck or confused. This enables organisations to evolve service models, reducing service delivery costs and increasing customer satisfaction.

Commonly, when people think of analytics, they think of a single tool like Google Analytics. As much as Google Analytics is a great tool, to get the best from data you need a broad range of tools to cover six specific jobs.

Data collection

To have data to analyse, you need some method to collect that data. The usual tools in the digital sphere include tracking code and tagging tools like Google Tag Manager.

Data storage

While it offers so many functions, the most important benefit of a tool like Google Analytics is it gives you somewhere to store your data. Alternative analytics projects might use other forms of data storage, such as a data warehouse.

Data interrogation

Another key function of Google Analytics is its interface through which you can interrogate the data. Other data stores use other interrogation methods such as R and SQL.

Predictive analysis

One thing that Google Analytics on its own does not do is provide any real predictive functionality. Commonly, this can be achieved by exporting data and building predictive functions in Excel or Google Sheets. However, Google also has other tools in its 360 suite, such as Google BigQuery, which are excellent tools for predictive analysis.

Data visualisation

Data visualisations in Google Analytics are good, to a point. The problem is that they are designed as a one-size-fits-all. This creates a problem when you have specific questions to answer quickly, especially where an answer relies on multiple data sources. The solution is to use a data visualisation tool, such as Google Data Studio or Tableau.

Data insight

The single most important tool in the analytics toolbox is the tool that gives you the actual data insight. Data insight is made from the stories that the data can tell you about the patterns in data and therefore tell you why things have changed. In almost all cases, this is where the human brain comes in.

Where to start with analytics?

The first thing an organisation must do is create a blue sky vision for how the organisation could use data.

Without thinking about any predefined metrics or KPIs, answer the following questions: What does the organisation really need to know to make decisions? How would you know when you have been successful? How quickly do you need to react to bad performance? How do you know your data is accurate? How do you know your data is safe?

Once you have the answers to these questions, the organisation must audit its analytics maturity. How does it use data? How does it store data? What data protection policies are in place? What is the reporting governance model like? Is there real time reporting? Is there predictive analysis? Does its use of data meet the blue sky vision you have painted?

To be considered mature, you need have a clear and well-defined analytics service embedded throughout your organisation which ticks all of these boxes.

From all the information gathered in these processes, the organisation can begin to build a plan in which it can use analytics effectively. It’s important to stress here that analytics maturity is not something that can be taken from low to high in one giant stride. Rather, it is an iterative process, where the basic data needs must be serviced first, and then additional layers of complexity can be added later.

Analytics maturity evolves as the business needs of an organisation evolve. It is a never-ending process. Therefore, each step along the path of digital maturity should be bite-size, specific and manageable. The road to analytics maturity should continue step-by-step.

While there are many reasons why analytics is important for your business, establishing how to use it effectively is key to not only reporting on performance thus far, but predicting what lies ahead. If you’d like to find out more about how Storm can help you audit your analytics capabilities, get in touch now.