Here are some things I wish I had known over the past few years, when it comes to Business Intelligence.
First, what is business intelligence?
This is how Business Intelligence (or short said “BI”) are being defined on Wikipedia:
Business intelligence (BI) comprises the strategies and technologies used by enterprises for the data analysis of business information. BI technologies provide historical, current, and predictive views of business operations. Common functions of business intelligence technologies include reporting, online analytical processing, analytics, dashboard development, data mining, process mining, complex event processing, business performance management, benchmarking, text mining, predictive analytics, and prescriptive analytics. BI technologies can handle large amounts of structured and sometimes unstructured data to help identify, develop, and otherwise create new strategic business opportunities. They aim to allow for the easy interpretation of these big data. Identifying new opportunities and implementing an effective strategy based on insights can provide businesses with a competitive market advantage and long-term stability.
The definition is pretty broad, but in most of the companies I have seen, it’s mainly the team in charge of designing, building and updating the dashboards and reporting suite, should reports be for C-level of simple managers.
Should Business Intelligence be centralized or decentralized?
I actually have seen both these two approach of BI being implemented.
Sometimes, withing the same company, management to switch from one to the other over time.
So, on the one hand, liberty is granted to build reports. It can be at a department level – imagine that the marketing department has its own BI team, designing reports and dashboards tailor-made for the function as topic matter experts or it can also be at a region level; let’s say that Europe, Asia or North America regions have their own BI capacities and dedicated people to report on their respective P&L.
It can also be that the BI grows organically since every analyst has the right to publish reports.
On the other hand, I also have seen the other approach of centralizing all the reports; having a small set of highly-skilled people in charge of building the reports – while stopping all local teams to build any report or dashboard.
So what’s work best?
Honestly, neither one or another, but a clever combination of the both.
Business Intelligence is about how you read the business. Teams align and managers make decisions based on numbers. If, on a given topic, two reports tell at 90% the same story but the nuances are not the same (e.g. the KPI definition is not 100% align between two reports), it opens the gate to endless discussions, politics and ultimately waste of energy and churn among the teams. We all have been through meetings where managers question what’s been included or not.
From experience, the most important KPIs and metrics need to be handled by a global and centralized team, while locals only can publish reports on topics where the actual scope or definition is 100% local.
Global execution should be the rule (~ 90%) while local execution needs to be authorized but exceptional.
Some example of it.
|Topic||Global / Local|
|Measuring the # of customers acquired on a given period? Global||Global|
|Measuring the financial metrics||Global|
|Impact of a law modification on a country on financials||Local since only the local Business Unit has the knowledge|
|Marketing metrics||Global (to avoid discussion)|
|New local product||Local at start – then Global once scaled|
Should we use only one software or should several be allowed?
Business Intelligence software landscape looks saturated these days: Qlik (Qlikview and Qliksense), Tableau Software, Power BI, Cognos, SAS, Looker, …
Honestly, – and I’m speaking only from what I saw, only if there’s a budget constraint reason, having multiple licenses in parallel doesn’t hurt – and each solution, having its pros and cons, should be considered.
Qlikview, having its own compression data storage format can be good for data visualization requiring a complex and heavy data structure.
Tableau software however generates very clean views and has a great data visualization engine offering more design possibilities.
PowerBI embeds super easily with Excel – being itself a Microsoft product- Finance teams will like it.
The most important is to get a unique official portal, curating the most important reports, regardless with which software they’ve been built with.
There are solutions on the market, making it possible to curate the reports, and offer like a meta search to all the users across the company. One cheaper alternative could be to develop such a website internally, storing the links to the dashboards in a MySQL database and building a simple search.
BI is nothing if datamarts don’t exist
Like one exec said to me once: “S**it in, S**it out”. He was right.
Business Intelligence can’t exist without good and cleaned data layers. Better said: Good Business Intelligence can’t exist without good data.
The role of the reports produced by BI are meant to become source of truth for all the teams, from the employee to the senior management of the company. The required step wich will be to partner with the company’s data team to build official datamarts which will operate as source of the truth company-wide. Building reports on them becomes actually an easy piece of the process.
This means that the data engineering team and the business intelligence should work hands in hands – and share the same data definitions, databases and agree on common processes and roadmap.
If a table breaks, BI needs to know immediately about it – since this has direct consequences on the existing reports. BI needs also to specify their needs in terms of data structure. Data structures and datamarts supporting the reports should be 100% owners by the data engineering team. BI should never maintain a table in order to keep the two tasks separate and avoid confusion.
To wrap it up…
There’s no perfect solution, it also depends a lot on the culture of the company and its maturity when it comes to data. It’s a journey and BI should also be flexible enough to meet business constrains: new product lines to monitor or stop, new regional cut, new leadership team for instance.
If you have any question, don’t hesitate to contact me 🙂