At the risk of being the unpopular consultant, I'm going to make a recommendation that no one wants to hear. Please. Stop spending money on visualization software.
If you Google "Top 10 BI Software" or go to a trade show, you'll see all sorts of visualization tools that promise to sing, dance, and make your dreams come true. It'll make dashboards, KPI's, and pretty charts--maybe even warn you if you drop below Safety Stock. Wait, put the credit card down, and please stop buying visualization software. While it is fun believing that a shiny new software can solve your data problems, you probably already own a visualization tool that will meet 90% of your reporting and analytics needs! Don't believe me? I'll bet if you asked an analyst to give you numbers about <insert your business metric here>, after a few hours, they'd come back to you with an Excel Spreadsheet or Pivot Chart or two. Turn around and ask the Finance Department for an Income Statement report broken out by Department and Region with measures for Current Year, Previous Year and %Growth. Again, I bet you get another spreadsheet. And why not? No matter how unsexy it seems, Excel is the de facto tool for business analysis, and it will probably meet 90% of your reporting needs. The real question though, is how much sweating did the analyst have to go through to get you those spreadsheets? If you look at the formulas and see a slew of SUMIF()s and VLOOKUPS() how nervous are you about presenting to the bank, auditors or in meeting rooms? Only way it could be worse is if you saw a bunch of static numbers! Truthfully, this is where BI comes to play. Although our eyes may start glazing over when you start talking about SSIS, ETL, star schemas, and multidimensional models, the fact of the matter is, any visualization tool will perform better when pointed at a well-designed Data Warehouse or Cube. In fact, if you ask the vendor giving you your Visualization Software demo, I’ll be they’ll tell you that data is coming out of a sanitized data warehouse, cube or data model that emulates a star schema. If they’re using none of these tools they’re assuming that your ERP contains clean data. <Insert skeptical snerk here>. So what exactly is Data Warehousing all about? Well, the Data Warehouse is actually the end of a tumultuous story of data access and extraction from your various source databases, Point of Sales systems, Google Analytics, Bloomberg CSV's, customs spreadsheets, or web-based Census Bureau data). Once extracted, data undergoes rigorous application of business logic and queries to conform, validate, and clean all the disparate data, then finally a transformation process forges the cleaned data into a Data Model that consists navigable dimensions and fact measurements. Voila. Your Data Warehouse. Sounds dry and possibly unsexy? Consider the alternative. Earlier this year I worked with a client who wanted a ranked dashboard that would show the amount and volume of deals that salespersons were bringing through the door. The first draft of the report took less than an hour to build, but it was immediately sent back for revision. "That lead conversion number is almost triple our average," I was told. Further investigation revealed that Status could waffle between Lead and Conversion multiple times based on a combination of factors including: data entry problems, whether the data was collected via web forms or hand written forms, or if specific intermediary statuses were made or missed. My report needed Olympic-quality gymnastics to extract a reasonable measurement at the cost of report performance. Suffice to say, what should have been an easy report ended up taking more time than expected because there wasn't one source of clean data, and I loathe to think what would happen if the client wanted the report rolled up by region, or presented in a different layout. Every client this story of having paid a consultant an ungodly amount of money for a report that we’re only marginally comfortable with and cannot be modified to present the data a slightly different way. So what’s the solution? A new visualization tools? No! Invest in a clean data warehouse that handles all the data gymnastics for you! Don't get lured in by a vendor's siren's song of sexy graphics, KPI's and dashboards if you don't have a solid data warehouse already. Instead, I would recommend looking around for a Data Warehouse automation tool. Many of my client have enjoyed success working with the Jet Enterprise data warehouse automation tool, http://jetreports.com/products/jet-enterprise.php. And I’d be thrilled to work with you to develop data warehouses either from scratch or customize one from the library of pre-built projects. Once that’s done, THEN we can talk about sexy visualization tools.
0 Comments
Your comment will be posted after it is approved.
Leave a Reply. |
Categories
All
|