PivotÂ is a software application fromÂ Microsoft Live LabsÂ that allows users to interact with and search large amounts of data. It is based on Microsoft’sÂ Seadragon.Â It’s been described as allowing users to view the web as a web rather than as isolated pages.
Here is a catalogue of a FMCG products. We are very proud of it.
You are welcome to play with the “business intelligence” PivotÂ here.
The watch pivot can be accessedÂ here.
You will see that some watches are “mixed up” (the picture does not correspond to the info. The reason is that this is just a demo, an example. We are not in the business of helping people find watches ;). We are in the business of helping brands use this great tool.
This TED video explains what Pivot does much better than we could ever do.
There still are many circumstances in which the main problem is to â€œget holdâ€ of the data we need and find a way to analyze it effectively.
However it is getting easier to find data in â€œrawâ€ form (see for instance Microsoftâ€™s Project Dallas), and it also getting easier to process that data in your PC without special IT knowledge (Microsoft Powerpivot add-in to Excel, a good example of so-called â€œself serviceâ€ business intelligence, lets you do things you would not have dreamed of 2 years ago).
Once you have the data, lots of date, in the form you want, the problem becomes finding a powerful way to visualize it, to make sense of it.
If what you are working on is something where you have plenty of data items you want to see â€œtogetherâ€, but you are also interested in â€œwhat the single data item is doingâ€, then Pivot is your tool.
For instance, to really understand what is happening to sales, you need to see the general picture, then look at a few items, and then go back to the “mass” (often slicing it under a different angle).
This is an example of a business intelligence Pivot for an imaginary company
You see all the products ordered by sales. Green products sell the most. Red products sell the least.
You can play with the data, for example filtering by product.
It looks as if t-shirts are doing very well.
Since natively Pivot knows how to “count”, but not how to “add”, we have developed a small add on that can, for instance, “adds” all the sales of a given view of the data.
What is wrong with the red shirt? It’s sales are less than 10.000 Euros, and it has a very low margin (the single Euro coin).
I can zoom further, and discover â€“ for instance â€“ how this item is doing this year relative to the other products of its category.
And so on. With a little tweak, with Pivot you can not only filter data, but also slice and dice across multiple dimensions.
What if you are interested in buying a watch, but you donâ€™t know which one to choose?
If you want to buy a watch most likely it’s not to â€œknow the timeâ€, but because there are one of more brands out there that have captured your attention.
Still, a bit of help in cutting down the possible choices could come handyâ€¦
Here are 7300 watches.
I love Breitling (254).
Interested in a ladyâ€™s watch (26)?
In gold (4)?
I like the one to the left.
Lets talk about something else.
Say you are the marketing manager of a successful and loved brand, that is also successful on Facebook.
You have a Facebook page, like everybody else.
Unlike everybody else, you have hundreds of thousands of fans. The page posts often and get hundreds of comments per day from its fans.
So what are you supposed to do?
Are you going to read 500 comments every day? No. The Media agency is better at doing that. Still, you would like to know what is happening.
The problem is that it is difficult to define â€œwhat is happeningâ€.
It is not “how many comments I got today?”.
A graph that shows sales going up and down over time is important. A graph that shows number of comments going up and down over time is much less important. Some page posts engage more, and other less, but the former are not necessarily â€œbetterâ€ than the latter.
It might not even be “what is the % of positive comments?”.
Loved brands are loved, and generally get only a handful of “negative” comments – that is it often safe to ignore, since the just represent background noise (you can’t be loved by 10% of the people right?)
“What is happening” should mean something like “how much/well is the page contributing to whatever goals the brand set for it’s Facebook page?”
Letâ€™s give Pivot a try.
Here are the comments from the months of March, April and May taken for a fairly important Facebook page. About 7000 of them.
A lot of â€œfansâ€ are posting â€œspam-likeâ€ content, â€œusingâ€ the page to promote their own wares.
If you filter only the “spam” comments, you see that they are growing and represent nearly half of the total number of comments (3000). Spammers seem to be equally divided in boys (blue icon) and girls (pink icon).
Who are these â€œspammersâ€?
For instance, lets look at the top 4 female “offenders” (names not shown). About 50 posts each. Almost 1 per day.
What are they saying exactly? One is aggressively promoting her collection of bike pictures. Which is OK, however a brand might or might not want its hard-earned Facebook real estate to be used to promote bike pictures.