Checklist
I myself have once examined what the reason for this could be and have compiled a preliminary list:
- The user does not realize that he needs information (awareness);
- The user does not want to use (want) information (facts);
- The user does not know that the information is stored (knowledge);
- The user does not know where the information is stored (knowledge; location is unknown);
- The user does not know how to formulate his question very well (skills; does not know how to get to the location);
- The user does not have time to find the information (time);
- The user sees the information (physical) but cannot place it in context or interpret it very well (cognitive).
Our huge appetite for good information is understood by Google
Google has understood like no other how we people want to deal with our huge appetite for good information. The company offers ease of operation, fast results, and excellent guidance (help, suggestions, alternatives; automatically correcting spelling errors). I personally find searching with Google a pleasant experience.
Google cannot do a lot for the first two points and the last point from the above list. However, it can be a great help with the other four.
Connect Google directly to your data warehouse
I want to make a plea for connecting Google ‘directly’ to your data warehouse and business intelligence system. This can be at the report level (document), by chart or table, or perhaps even by fact. The more detailed information Google can disclose, the better the results can be. Furthermore, nowadays many of the better (enterprise) BI tools can handle Google’s requirements pretty well (for example Power BI and Qlik). I have only come across a few implementations where this functionality is applied and that is a pity.
Visitor numbers will rise
You will be surprised to see that the visitor numbers of your data warehouse increase significantly, as well as the number of hits. Metadata then becomes more important than ever, otherwise, we cannot apply a number of important techniques and principles for search engine optimization. Think about keywords, description, and title. However, we will also need to provide the appropriate tags for each record and attribute in the data warehouse itself, both at a detailed level and at an aggregate level (scripts can be written for this).
What profit can you gain with this?
You might be wondering what the profit of this could be. A few examples may clarify this.
- A manager types the following into Google: “what is the profit for 2009″. Google gives the answer immediately, perhaps even displaying the accompanying chart, including a link to the report or dashboard.
- An employee enters: “what is the turnover of customer x”. Google responds with a link to a (directly) generated page on which a table is displayed with the turnover of customer x, for example, over the past 13 months. Simultaneously Google displays links to external news stories that contain the name of the client.
The big advantage is that the user does not need to log into a system; almost never has to wait long for an answer; he is helped in his search for relevant information from that time; He is presented with alternatives; and external and internal management information can be combined.
Asking computers questions with natural language
The problem in connecting Google to the data warehouse no longer lies so much in the fact that Google should be able to understand your question. Of course, natural language is a phenomenon that is not easily captured in a structure. However, there has been so much research done on it that it seems to be only a matter of time before we can ask computers reasonably good questions while using natural language.
What is still a major challenge?
Where I think there is still a really big challenge, is in the converting of ‘all’ information in the data warehouse to indexed web pages with images. These web pages have to be strongly interlinked with each other and neatly laid out, just as it is good practice to do this on ‘normal’ web pages. The user can click through to related issues directly from the page and be quickly informed. After all, we want to be able to serve the user properly on the data warehouse website.
Will we continue to muddle through with pixel-perfect reports?
The big question is whether the market is going in this direction. Or will we continue to mess about with building inflexible pixel-perfect reports and rigid dashboards? Time will tell, but the dynamics can sometimes become so powerful that only ease of use and speed can save us.
Daan van Beek MSc, speaker, trainer and author of the BI & AI book 'Data Science for Decision-Makers and Data Professionals'. He is the CEO of the Passionned Group and you can follow Daan on Twitter.
