Analyse Your Business Data Analytics Needs

Nowadays many people talks about data analytics, data-driven decision making, big data, business intelligence, etc. And if we ask them: "Do you have any great examples about how your organisation leverages data to drive business value?" Their answers can be grouped into three types:

1. Success: "Yes, I have some success business cases, for example ..."
2. To be confirmed: "My company is capturing a lot of data, and generating many reports, and I believe we're driving better business decisions from data..."
3. Not started yet: "No, actually I am not sure where to start..."

If you're type 2 or 3, and you want to move to type 1, this article might be helpful, especially when:

• There are many data analytics requirements (and opportunities);
• But you have only limited resources (time, budget, people);
• However, you still want to maximise the ROI (return on investment) of your data analytics resources

Let's get started!

When a requirement comes to you, it could be a phone call or an email like: "Hey Ivan, we're running an online product purchasing journey, people can apply for a free trial via 3 different channels, as long as they register our offline campaign...Can you help me capture the data so that we can closely monitor the performance of the project?"

Pause.

Instead of saying: "Sure, what do you want to analyse?" I suggest you start with analysing this data analytics need.

The key is to gain more context.

And there are in three dimensions we can consider: Data Richness, Lifecycle and Scope.

Data Richness

There are five Data Richness types:

Type 1: Blackhole

• There is no data at all

Type 2: Desert

• Limited data
• Limited possibilities

Type 3: Silo

• Isolated data
• Comprehensive data
• Can only be accessed by a team or a department

Type 4: Lighthouse

• Limited data
• But data is consistent so it can be understood by wider groups in the organisation
• And data is accessible by wider groups as well

Type 5: City

• Comprehensive data
• Consistent data
• However, huge resources is needed to build a city

To summarise:

Now you want to consider:

• About the data we currently have in this project, how "rich" is it right now?
• Where do you want to go? Silo, Lighthouse or City? Remember, City is not always your best choice because it's "expensive".

In order to make better decision, you'll need another dimension: Lifecycle

Or change the way of asking: "How old is this project/channel?"

I usually breakdown lifecycle into three types:

Type 1: New Born Baby

• Fresh and new
• Getting attention and resources
• Flexibility for change
• Important stage for foundation

Type 2: Adult

• Serving lots of customers currently
• Can be both valuable and risky to make data capturing changes

Type 3: Elderly

• Maybe too "old" for big "surgery"
• Not very valuable to capturing comprehensive data from it
• But having a lot of "memories" that be helpful as "lessons learned"

Obviously, if you realise that the data requirement is about capturing a button click data from a very old platform that is going to be demised within 1 month. And the effort of adding tracking code to this button can take you 2 weeks. Then you'll probably never start accept this data capturing requirement at the first beginning, because it's not very valuable.

Here comes the last dimension I want to introduce: Scope.

Same as Lifecycle, Scope can be broken down into three types:

Type 1: Whole Body

• The data requirement for one whole website or mobile app
• The data relevant to one product across all channels: website, app, offline, etc.

Type 2: Arms and Legs

• The data of a purchase journey in your website
• The data of all pages that provides product related information
• All the login and log off related data in your mobile app

Type 3: Fingers

• Button click data on one of your promotional pages
• Tracking the paid search traffic from Google only

Here you can see, depending on different data analytics Scope, the time and effort needed varies. For example, if the data requirement is about all purchasing journey fallout analysis in the whole website, then this can be considered as Whole Body, because in order to get this kind of scope of data, you'll need to get in touch with developers working on all purchasing journeys on your website. And a standard way of data capturing strategy will be required.

On the other hand, if the requirement is just about analysing the button clicking rate on one promotional page, you can consider do it within 1 week. However, you might also want to consider take a step back, asking the requestor: "Hey, maybe you will like to know all the CTA button click-through rate on all promotional pages, breaking down by traffic sources and product types?" In this case, you started with a "Fingers" scope request, but you actively turn it into an "Arms and Legs" scope, but the time and effort you spend in combine will be much less than doing several "Fingers" separately.

Now let's put this into practice!

Case one:

A newly released mobile app product owner comes to you asking for the data requirement about monthly active users, their top engagement types and also wants you to support cohort analysis.

Data Richness: Blackhole

Lifecycle: New Born Baby

Scope: Whole body

Decision: Plan a comprehensive data capturing and analysing strategy for it. Because for such a new born baby, here comes the "now or never" opportunity to build a "city" on it. As long as resources allows you, go spend some time on it!

Case two:

Marketing campaign manager is curious about how many visitors are using the promotion checking tool on your website. She comes to you asking for the button click data on the tool page. You checked the page, there are some basic page-level data available (pageviews, visits, etc). However, you heard from the marketing campaign manager that they're going to launch a new interface of the tool next month.

Data Richness: Desert

Lifecycle: Elderly

Scope: Fingers

Decision: Ask more about why she's asking this data, and is it relevant to the new version of the tool? If not, just give her some basic page-level data to understand how many visitors and pageviews the tool is receiving in last month. Without spending too much additional time on the tool.

Case three:

A operation specialist comes to you wondering if you can provide some data around the online booking system, they're seeing much less booking in the past few weeks. You checked the booking system and realise that there's no much digital data on this booking system, however the customer data captured in the backend are quite comprehensive.

Data Rickness: Silo

Lifecycle: Adult

Scope: Arms and Legs

Decision: First of all, you know that this booking system is critical because it serves hundreds of bookings per day, which is driving revenue for your business. However, you also aware that the system has been there for a while, and it can be hard to make code-level changes for digital user behavioral data capturing and real-time marketing automation. So you will need to do further requirement analytics in order to make the decision:

A. If there's not enough resources from operation team to do code-level change, maybe you'll just dig deep into backend data and trying to get the answer for their business question.

B. If there is enough resources to do code-level change, and there're even more marketing budget to be spent on driving traffic to this booking system, then it's worthy of re-do data capturing part of this whole system, bringing it from Silo to Lighthouse or even City, depending on more context you gain.

Ok, I have shared my ideas about analysing your analytics requirements, now it's your turn to put this method into practice.

Please feel free to leave comments and let me know your thought about this!