Actual article date: Jul 7, 2015

The other day I was on panel session at the Singapore MICE Forum from which we saw data management as a key challenge for organisers. On the same day, I was walking back home from a fulfilling time at SMF when I chanced upon this:

Enterprises have historically spent far too little time thinking about what data they should be collecting and how they should be collecting it. Instead of spear fishing, they’ve taken to trawling the data ocean, collecting untold amounts of junk without any forethought or structure. Deferring these hard decisions has resulted in data science teams in large enterprises spending the majority of their time cleaning, processing and structuring data with manual and semi-automated methods.

Data collection remains a major issue for companies. However, there is no easy shortcut to it. Time and effort have to be invested to reach a state of good data management. This is because the reality is that..

“You have to start with a very basic idea: Data is super messy, and data cleanup will always be literally 80 percent of the work. In other words, data is the problem.”
- DJ Patil, Chief Data Scientist of the White House

At Jublia we handle data on a daily basis for clients as well as using data for us to do what we are good at: business matching. Here’s are three main areas where we have learnt to ensure quality data management:

Know the types of data you are collecting

The other panelist, Sherrif Karamat, from PCMA who leads his organisation’s digital strategy briefly touched on the types of data that we can collect. There are two main categories, 1) Stream data and 2) Static data.

Stream data are data that is determined by a user’s action and most often then not, his/her needs. Example includes clicks on an app, person’s search entries etc.
The benefits here are:
- The data can be more genuine when compared to static data.
- Data can be collected at scale and more representative of your audience.
The limitations here are:
- The data that is collected is huge in size and cannot be analysed manually via Excel.
- Data needs to be processed/crunched or put in reference with other datasets. It is mostly useless by itself.

Static data are data that is collected via forms. This include surveys and in events, registration forms.
The benefits here are:
- These data can be used in it’s raw format without having to process/crunch it.
The limitations here are:
- People may not provide genuine comments thus it may be misleading to what you are analysing.
- People may not know what they want. They just indicate what they think they want.

Collect data well

Collecting data is a tricky business, simply because we have so so many channels to collect data from. So how would we know what type of data will be useful for us to analyse? We have to look at what is our business aims and what type of data is fundamental to our needs.

Since Jublia is deeply involved in the business matching/networking process of clients’ events, we only collect data from channels that will be highly relevant for us to analyse further and help clients improve their events. And we know that collecting data is not as easy as it seems.

To collect quality data that is fit for analysis, we need to understand the channels that the data is streaming in. For instance, Jublia’s tools that provide our clients’ end customers (delegates, exhibitors, sponsors) have been built for performance with a streamlined user experience and clear benefits for them to use it. This improves the rate of adoption of our tool and as a larger percentage of a group is actively using the channels, the data streaming in is highly relevant and most importantly, representative of the cohort.

Simply put, you can’t have 10% of the people at an event downloading your event app and then claim that sentiments can be measured at the event.

The fact that “garbage in, garbage out” is still highly relevant even with the numerous technologies presented around us means that we need to know the quality of the data that we are collecting.

Use the right data framework

Having the right framework structure that you can feed your data in is very important. It is only with the right structure of your data can you activate them for further analytics and insights.

As an example, our recently revamped data structure has allowed us to analyse events much more deeply by having highly specific categories of different groups in an event, as well as attributes that can be assigned to each group. It enabled us to effectively measure metrics in different groups and how they perform. Furthermore, it enabled us to seamlessly overlay datasets with each other to further generate even useful insights.

One of the most intriguing sight I have seen is when I walk through some offices of enterprises and see rows of folders and reports filed in physical copies by their events year-on-year. They are most likely archived and referenced back once in a while. Those data could be really valuable if it can be stored in an effective format to ensure that the data can be easily activated for insights.

‍

If you have a any questions or doubts on the above, please feel free to speak to me. I am happy to share more and explain further.

— -

The quotes referenced in this article comes from TechCrunch: Enterprises Don’t have Big Data, They just have Bad Data

‍