In an ideal world, it would only be one. But analysing data is more difficult than that. When looking to analyse data there are three things you need to consider.
The first is extracting the data. This should be a straightforward task, but the quantity of data you have collected means it can be time-consuming. The next thing you want to do is transform this data into something which is going to be useful to you. This means you want the data to be displayed clearly and easy to consume. From this, you will load it into an exploration tool to then report on it.
Until now, this is how most businesses had accepted they would have to analyse data.
However, at Webtrends Optimize, our goal has been to create a seamless experience when analysts are interpreting data.
This is why we have created our brand-new reporting engine, Discovery. It allows you to analyse experiments in more detail, observe data from different viewpoints and discover intelligent insights faster than ever before.
Why is the Discovery Tool necessary?
Testing tools have traditionally been limiting. Part of this is due to insights often being slow to come by. For example, in most testing tools, if you want to understand the performance of your desktop, tablet, and mobile, you must load the UI in multiple browser tabs and filter each tab to each device.
To compare, you will likely have to manually copy the data cell by cell, into excel and then begin dissecting from there. This isn’t very efficient as it makes it hard to scale (if you wanted this to be 300 cities rather than 3 devices, it would be very tricky) and it relies on making no errors to get accurate data (if you input one number in incorrectly, it could ruin all your results).
Moreover, it is harder to guarantee the data you are receiving is accurate. For example, you may have built a test and QA’d it on Chrome and Safari. You run it for desktop, and find it isn’t particularly impactful. You learn what you can and move on to the next one. However, what if Chrome and Safari were at +5% for conversion, but IE and Edge were at -40%?
What if some countries don’t gel with the experiences you've built? Or there are device types you didn’t account for in the design?
Bots could regularly crawl your site, enter your tests, and simulate real browsing through headless browsers. Recognising them is great - but removing them from the data is crucial. Bad outcomes are easy to come by, and having dirty datasets isn't helpful. This will slow things down hugely and often waste time for all involved.
What are the limitations?
In simple terms, the limitations are time and skill. It is quite a slow and consuming task to sift through and understand the data.
However, it also requires a certain type of skill. You need someone to be skilled with a programming language, like Python or R to write your own stats calculations.
This is where we come in…
At Webtrends Optimize, our goal has been to make this a more seamless experience. Our new Discovery reporting engine gives you answers in a few clicks, compared to before which could have taken hours to achieve the same results. You can do 100s of explorations, whereas it previously would have taken weeks to see the same success.
Discovery solves these issues by giving you the power over how you want to view the data. We provide you with the tools to apply dimensions to analyse and ask questions about your data.
As described by Sandeep Shah, Product Director at Webtrends Optimize, “We’ve built something I’ve not seen in any experimentation tool to date – by nudging where the custom dimension is inserted, your view changes completely”.
Previously, analysts were using multiple tools. But with Discovery, you can find what you are looking for in just a few clicks. Allowing for more insights, providing more learning and overall, more successful programs.
