Click here to monitor SSC

From DBA to Data Professional and back again

Published 31 January 2013 5:04 pm

One of my periodic tasks in the department I lead, “IT Data Services”, is to produce summary reports on the current state of my team’s projects, as well as statistics on our SQL Server infrastructure. For efficiency, I normally just write a quick query to pull the data from my DBA repository database, throw the results into an Excel Pivot Table, and “fancy it up a bit”. In the past few weeks, however, I have been working on a Business Intelligence (BI) solution with one of our development teams, so I thought I would go an extra step, and woo my boss with an interactive PowerView report. This is how I almost lost my job.

My emotional journey into the Microsoft BI Stack began by carefully installing SQL Server 2012 and SharePoint, and marrying the happy couple, with SSRS and SSAS in attendance. I choked back tears when I finally had all of the requisite and interdependent services and authentication modes up and running. I cheered when I saw that I could deploy and schedule a SSRS report in SharePoint. I high-fived myself (I generally work in a dark and lonely office) when Silverlight spun up to show me the galleries of Excel PowerPivot workbooks. I picked up the phone and called my mom to thank her for giving birth to me, just so I could witness the elegant dance of PowerView with my tabular data. Finally, I got on my knees and prayed that at least one of my 453 Twitter followers would share my joyous tweet, on seeing my PowerView report exported to a PowerPoint slide.

I sat back and reflected on the journey as a whole. It had not been without its problems, prime among them being numerous issues with authentication, data not being recognized as being in a proper date/time format across all technologies, and PowerPoint failing to beckon me to “Click to Interact” in the desired manner.

However, problems, lead to solutions and solutions lead to knowledge. I felt I was empowered with a new BI insight. Prior to my experience of building, configuring, developing, administering and analyzing data in “The Stack”, I had been but a DBA. Now I was a Data Professional, SME extraordinaire.

I was confident. So confident, in fact, that when my boss asked for a few reports, I assumed that he’d be so elated when he saw the new, interactive report in PowerPoint that he’d probably ask me to put away my DBA hat forever and become Chief PowerPoint Officer for our company.

I breezed through the query to pull the desired data, pumped it all into PowerPivot, published it to SharePoint, dragged-dropped, sliced-diced and charted. I emailed it over and casually strolled into his office to watch his reaction. I thought of shutting the door so that his boisterous enthusiasm would not disturb the others, but ditched the idea.

“What does this mean?” he said, pointing to a column header.

“Oh, that is just a generic name. I can clean that up!” I replied airily.

“OK, so how can this be right? I mean, how is this even possible?” he continued. “We have 34 million users for one database?”

“Gah, that’s in Gigabytes instead of Terabytes…” I stammered, attempting a crazy diversion into user counts measured in bytes. He didn’t bite. “Hey, look at that chart!” I tried.

“And how many servers do we have for this process? This says we have six. Is that right?”

I abandoned all faith in the accuracy of the report and began counting on my fingers from memory. “I need to double check, but I believe it is six, yes.”

Would you bet your job on it?”

There were seven servers for that process.

Dazzled and seduced by the new world of the data professional with its ever-evolving tools and techniques to manipulate and view data, I had broken an important rule: “Never demo anything to your boss.” As a golden addendum to that rule, “never, ever, demo anything to your boss without first making sure the data is accurate.”

I heard about another new technology called Data Quality Services. That is next on my confidence re-boosting list of technologies to master.

13 Responses to “From DBA to Data Professional and back again”

  1. callcopse says:

    Ouch, nice tale of technological hubris. PowerPivot is some cool stuff, too cool for me to use. When I saw the word SharePoint though, that’s when I knew this would not end well.

  2. sirwilliam says:

    Nice story. I allways include disclaimers when i present data, to remind my boss that (1) data quality is never 100% and (2) not all relevant conditions / filters are known active. Most bugs are in the requirements.
    So information needs to be peer-reviewed, that’s my experience too. And every time I have the feeling that I can lose prestige.

  3. SinisterPenguin says:

    I feel your pain!

    You do a cool new report, take a look at the data, do a few tests decide yep that looks GOOD.

    Then those darn users come along & immediately find some weird edge case you hadn’t considered & the whole thing either explodes or throws out nonsense data.

    You have to learn to check, double check & check again any kind of public/boss/team reports.

    It’s all this validation, checking & re-checking that makes BI work sooooo painful.

    Lets go back to our basements & leave the BI to the devs…

  4. jerryhung says:

    This was a funny read, thanks for that

    I think bosses should be on the “minimum” need to know basis, and employees should stick to what has worked :D

    or, get Microsoft sales rep to talk to your boss first to get the buy in (and pay) on these fancy BI technologies

    While BI is great on MSFT side, I think for reporting/dashboard, there are many other “better” products that management use (speaking from experience, where I always wonder why we have 5+ reporting solutions, from Business Objects, QlikView, etc..)

  5. Keith Rowley says:

    I love working for a place like you describe where I have the freedom to do some of my own experimenting with some of my time without having to clear the whole project with my boss first.

    Remembering to double check results before presenting them is key true. But you should be thankful your boss does not try to micro manage your every step. I know I am.

  6. Sergio E. says:

    What a story!

    I know how you feel and the whole process of joy and the the big fall to a dark abism of terror when the presentation fails…

    And as usual, it’s an new invaluable lesson.

  7. BuggyFunBunny says:

    Giving the kiddies the keys to the liquor cabinet is never a wise idea. Nor is letting corner office Suits so neurosurgery.

  8. paschott says:

    Great story. I love the ideas behind PowerView and some of the MS self-service BI stuff, but our challenge is getting it to work against Multi-tenant databases without giving our users access to another customer’s data. Add in the time it would take to research that particular bit of filtering and I don’t think we’ll be using PowerView anytime soon (unless someone has some pointers on restricting access based on who logs in from different organizations coming from a generic internet connection)

  9. willliebago says:

    It’s an awesome feeling getting all of the BI stack to work together seemlessly! Let us know how the Data Quality Services implementation goes :)

  10. kellym says:

    Dude, you have got to try Tableau for your analysis. I love MS SQL, but for investigative analysis and reporting? Not so much. With Tableau, you can connect directly to your db or cube and built a chart in seconds. With that extra time, you can make a few more and throw them onto a dashboard. Seriously – reduced my work time by 50%. Jen Underwood explains it well here:

  11. Natarshia says:

    I’m so glad I took time to go back and read this month’s newsletter. Your blog was great. Not because I have been there and done that, but because I had no clue of how this story was going to end. Most of the time we can guess what happens during a blog/story. This time there were so many ways this could have went it was like reading one of those choose your own ending story. I thought you was going to spend all that time on that infrastructure only to have your boss tell you he liked the old way. Or that even though it was self serve, the boss would still ask you to generate it and print it for him. Or that he was going to be mad all the work you did without approval. In any case, the moral that I would take with this story is that so many people think automating a process is easy. But it’s not. When you tell them how long it’s going to take to change a manual process, you feel like you are cheating them, but in reality, new things introduce possible errors. If something should take 2 hours, I always tell them it will take me a day. At least your boss was astute enough to even look at the data and question it and not just the format. The next time you can write a blog post call it ‘the right data, wrong format’. i love it when i get the data right but the format wrong. you know that hey it doesn’t make sense unless the column headers are bold bosses.

  12. pbyrne says:

    In my 20+ years of experience in the industry, I’ve found that there are two types of people: Those that make mistakes, and those that admit to making mistakes.
    In this case, Rodney gets bonus points for admitting to his mistakes on a public blog as well as my nomination for “SQL Man of the Year”.

Leave a Reply

Blog archive