Insight on Insight — Part 2 — All About Speed
A few weeks ago, I started my Insight on Insight series. (This is a supplement to the already informative Insight related blog that my colleague and friend, Michael Halbrook, Insight Guru Extraordinaire, is also writing).
And then I kind of took a step back — because, well, I had to work.
But I have been wanting to revisit this topic and so here we are, friends.
Today, I wanted to address one of the subjects I brought up in my last post on the subject.
What makes Insight a true differentiator in the marketplace?
Again. I am a vendor. I know that you just might take what I am saying as “suspect.”
However, I also want you to know — I am a terrible liar. I am very bad at the art of deception.
And I am terrible at Poker.
But here goes.
1) Speed to process — Insight can process massive amounts of data exponentially faster than traditional business intelligence solutions. Part of the speed that is gained is that the data schema is not cubed or tabular in nature, but rather, circular. Omniture works closely with its clients to size solutions in a manner in which the entire dataset can be REBUILT if needed in the course of 1 — 2 days.
I have worked with clients who will rebuild datasets over a given weekend to add new dimensions or incorporate new data feeds with 5+ terabytes of data.
Process starts on Friday night. Ends on Sunday.
Sometimes we hear that this is too slow. But this is from clients who started their dataset with 5 months of data and have now seen it grow to 15 months. Or have started with 50 custom dimensions and have now expanded that to 200 dimensions.
I guess the best analogy to that is — and this is just off the top of my head — you can buy a sportscar and expect great performance when you take care of it and follow the maintenance plan.
Or you can let it go, not bring it in for service and shove ten people in it and complain that the doors won’t close.
You wouldn’t do that.
The diagram above shows a simple view of the Dataset Schema within Insight. Within the demo schema above — there are five levels — Visitor, Engagement, Visit, Hit and Commerce. This dataset is an example of a multi-channel dataset which is combining customer interactions along with web activity.
(Most of the dataset is not seen as the image would be too long. Ping me if you are interested.)
All new dimensions, or categories, are associated to the individual level when they are added at the dataset. The lack of a tabular construct makes it possible to query across the dataset with greater flexibility than in a cubed environment.
2) Speed to change - One of the aspects of the software that make it so coveted by clients who know how to maximize it, is its ability to be flexible.
Like I mentioned above — as long as the data is fed into the system and can be associated through a primary key, it can be made into a new dimension in a relatively straightforward reprocess of the data.
In addition, the ability to save segments as dimensions and create your own metrics is an equally valuable proposition for clients whose business analysis needs can change with enormous fluidity based on the business demands of the day.
For more information on that, check out Mike’s piece on Segments in 5 Minutes.
3) Speed to answers — Traditionally, when making a query from a data warehourse, you submit your query.
And then you wait.
And then you get an answer.
So what do you do with that?
Usually, great analysis is iterative. It does not stop when you have the answer, because the answer usually leads to more questions. If you are an analyst, knowing that 733 is the answer to your question will cause you to ask more questions like how, why, from where?
As soon as you start a query in Insight — the system starts returning an answer. The immediate answer is an extrapolation based off of a local sample — but as the query reaches 100%, it becomes more and more exact.
What this allows you to do is start asking questions, and maybe some more questions as you see the fallout of trends. The sampling is highly accurate.
So you don’t have to take 10 coffee breaks waiting for your separate and very disparate questions if you are still working in the old order.
Which is a good thing, right?
I guess it is. Unless you really, really like coffee.
4) Speed to action — It goes without saying that having the ability to process so quickly and drive answers so efficiently enables your business to act more quickly and make relevant decisions to drive value.
I have worked with one large retail client who, prior to having Insight, had to scramble madly after the Black Friday rush to deliver findings and provide evaluation for C-level reporting. However, in trying to pull data from disparate systems and trying to merge reports — Black Friday analysis was not fully complete until … well …
Which is great.
Except at this point — the information was not really an impetus for change. Instead — it was solely that — information. Some level of insight could be extrapolated for future business decisions — and potentially to change behavior for NEXT Black Friday.
After implementing Insight and joining data that previously needed to be manipulated, massaged and interpreted for weeks prior to delivery, the company was able to make some dramatic changes.
A true paradigm shift.
What took months to analyze could now be delivered to the C-level team within days after Black Friday and with the ability to make changes to store inventory, website specials and holiday campaigns to directly influence the current holiday season.
So there you go. Just some Insight on some core foundations of Insight that elevates it beyond traditional reporting tools and makes it a true “game-changing” factor for companies who have embraced it.
More to come on my Insight on Insight series — Is Insight too Complex for Me?
You can follow Kiran on twitter @measureTHISgirl.