Toan Hoang

TL;DR / 3 Performance Considerations

A key aspect to delivering data analytics and visualization is the performance of your solution; performance can easily make or break a project, however, too often, we focus on a few narrow aspects. In this article, I want to share my three main performance considerations when planning and delivering business intelligence and data analytics solutions.

Note: this is not an article that is not focused on Tableau, my experiences are not solely Tableau, but I hope you will find this an interesting read.

1. Time-to-Deliver

Times change and technology is evolving at a freighting pace, and with that, so do our challenges. Client expectations around Time-to-Deliver are one of the things that have rapidly increased in the past decade. I come from an age where the time between a consumer requesting and receiving data could be weeks, months, or not at all, and, to add insult to injury, the request came with a time-consuming list of documents to be filled in and vague processes to be followed.

We live in a new age now, with short tweets, micro-articles, and self-service no longer a hard to achieve luxury, but a must-have for any organisations that strive to become data-driven. Gone are the days where each request comes with a 10-step process from data sourcing to report development, and here are the days where we can store near-infinite amounts of data, pre-process and optimize data, and have tools to make data a single click away. All provided we set up our data landscape appropriately, of course.

For your current project, ask yourself:

As a consultant, a major factor here is to understand the data landscape and data Capabilities (people, processes and tools). So let us break down this down our capabilities.

I have spent the majority of my career working across the data stack; from acquisition to storage, from data sources to end-users, and have always focused on improving my skills in each of these areas. Given that we live in a world where data volumes are increasing and expectations for Time-to-Deliver is shortening, so it is important to take this into consideration.

2. Time-to-Insight

This is where we look at the designs and, more importantly, measure the time between opening a dashboard (or report) and seeing insights; on my travels through various clients, I see far too often that outputs created for end-users are not created to deliver insights, but to deliver data. This is something that I push for. End-users having to go through various different menus, clicks, browsing around before they see their insights is really not a good thing.

When creating outputs for end users it is not about presenting KPIs, but ask them why they will open the dashboard, what questions do they have, and what they expect to see? This is a part of requirements capture, and far too often, it is left out in the world of fast and self-service data analytics, but so important it is…

For your current project, ask yourself:

Time is a precious commodity, and while you obviously want to have the information in front of you in a timely manner (Time-to-Deliver, which is a data engineering task), and you also want to ensure that the time to get insights (a visual design task) is also taken into account; working on both of these will be key to your project success.

3. Time-to-Action

The Time-to-Action is something that far too often is missing, and something that I have been increasingly focused on in my career. The Time- to-Action, for me, is the reason why Business Intelligence and Data Analytics projects exist. There is no point in having information if there is no action attached to it, or if there is an action, said action may take too long to perform.

Let me give you a real-world example, imagine a dashboard that is targeted towards your sales workforce and one that give a list of tasks that needed to be performed each morning. This sounds pretty simple, an automated task list leveraging a calendar and some insights to make sure that nothing slips through the net; the dashboard has insights, actions, and the list of quite comprehensive, so is pretty useful.

Now imagine when someone views the dashboard, and see a task to email a client about renewing their subscription. they then click on the email address and it opens Outlook, great, but now they have to go back and forth between the dashboard and his email to double-check that he wrote the numbers down correctly; being cautious, a total of 15 minutes to make sure the email is completed before sending. I am sure that you have seen instances like this in your various operations.

Now, we know that the consumer needs to send an email, and we know what content should go into the email, so why not open a Link to Outlook and embed all the required information there? If we did that, the user won’t be making sure that he copied the information across, but double-checking his message before sending. This is a simple, but hopefully illustrative example of how we can improve the Time-to-Action.

On your current project, ask yourself:

When looking at producing dashboards or reports, always think about the action and the Time-to-Action. The more you can reduce this, the better off you will be and the happier your clients will be.

Summary

I hope you have enjoyed reading this article as much as I have written it; I will be sharing more of my thoughts with you all, but this one is of particular interest to me.

Always remember to think about the three Ts, the Time-to-Deliver, Time-to-Insight and Time-to-Action, and work diligently to reduce these as much as you can as this will benefit everyone involved.

Lastly, please let me know what you think, share your thoughts below, or reach out to me on my various social media platforms. I would love to discuss.

Exit mobile version