Interview Questions for Data Warehouse Developers

Data warehouse developers are highly visible professionals who earn solid annual salaries – around $96,000 — because they’re responsible for using state-of-the-art development tools and technology to successfully deliver business intelligence information across the enterprise.

Dice Interview Qs IconMost candidates are screened over the phone or Internet by recruiters like Vince Brown, director of healthcare IT informatics and IT recruitment at ICON Staffing Network. “Candidates need to be truthful about what they know from the outset,” Brown says. “Otherwise, they’ll be hard-pressed to answer tough questions about their coding skills and their ability to lay out ETL mapping.”

We asked Brown to share some of his favorite screening questions for data warehouse developers.

Are you only an ETL developer or do you have other data warehousing experience?

  • What Most People Say: “I have strong, well-rounded data warehousing skills.” Then: “What’s that? You want to know what I’ve done outside of ETL mappings? Ah, I need to look at my resume.”
  • What You Should Say: “I have broad experience that extends beyond data extraction, transformation and loading. I have strong coding skills in SQL and Oracle. I’ve collaborated with data warehouse architects and I’ve helped configure several database and business intelligence tools. May I tell you about two of my recent projects?”
  • Why You Should Say It: The best way to showcase your skills and abilities is by providing an overview of recent projects. That way, the recruiter or hiring manager can get a sense for what you can do. Remember that every job is different and you may be a good fit for some positions but not others. (By the way, referring to your resume during the interview – which some people do — is a sure sign that something’s not right.)

How much experience do you have with Informatica products? On a scale of one to 10, rank your skills against other Informatica developers.

  • What Most People Say: “I’m a nine or a 10.”
  • What You Should Say: “I have three years’ experience and I’d rate myself a seven or an eight.”
  • Why You Should Say It:  Mastering Informatica is practically impossible because it’s a constantly evolving tool. “Most developers don’t even use the entire tool, so rating yourself a 9 or 10 raises a red flag about your honesty because nobody’s that good,” Brown says.

Are you familiar with Netezza? What type of processing does it use?

  • What Most People Say: “I don’t know what type of processing Netezza uses.”
  • What You Should Say: “Netezza uses massively parallel processing, or MPP. We use Netezza at my current company so we can search a number of databases in parallel.”
  • Why You Should Say It: Since the setup for MPP is more complicated and requires an advanced developer, Brown uses this question to separate rookies from veterans. More experienced developers usually explain how they’ve partitioned a common database or assigned work among several processors.

Have you done any data modeling work?  

  • What Most People Say: “Yes, I’ve used logical and physical models.”
  • What You Should Say: “I’ve created conceptual, logical and physical models using tools like CA ERwin, ModelRight and Visio. May I give you a couple of examples?”
  • Why You Should Say It: Examples are the best way to illustrate your hands-on experience with different data modeling techniques and tools. Some people have only observed them.

“The questions only get tougher from here,” Brown says. “You don’t need to ace every answer or match every single requirement, but you need to be honest about your strengths and shortcomings.”

Finally, Brown has experienced some candidates actually searching the Internet for answers during a phone or online interview. “I don’t want to hear typing during the interview,” he says. “It’s better to emphasize your strengths while painting an accurate and realistic picture of your experience with key development tools and processes.”

4 Responses to “Interview Questions for Data Warehouse Developers”


    Call me old-fashioned, but when DICE runs a series of articles about the “right” answers to a wide range of interview questions, aren’t you just helping unqualified liars beat out honest, qualified candidates? If someone wants to pretend to have used ERwin, they can download the demo version from CA’s site, buy a book on database design and study. You’ve bypassed all that trivial “learning” stuff and gone right to the answer key! Drawing pictures in ERwin and saying “conceptual,” “logical” and “physical” in the same sentence does not make you a data modeler.

    Case in point. I’ve never used ERwin. A business partner and I were evaluating an expensive insurance processing package and insisted on seeing the data model under non-disclosure. After much wrangling, the big day came. We were handed two paper packets that we had to surrender at the end of the meeting. My evaluation partner with the formal ERD experience flipped through the paper and nodded his approval. It took me about ten minutes and six pointed questions to establish that one model was a complete fabrication while the other model was obfuscated to avoid revealing anything meaningful about the software product. This guy is no doubt prospering as a data architect and I can’t get an interview because a software package gained market share and I wasted my time studying six different books on relational database modeling and wrote my data models, heaven forbid, on paper and in VISIO.

    Candidates that pretend to know more of the “nice-to-have” packages have a real advantage in the phone screen. HR can’t pick up on the visual cues given off by a liar over the phone and there’s no expectation that HR can do a technical screen. The result–“Oh my gosh! I found three sharp, energetic candidates with 4 years of experience who have done BI work and know ERwin, Informatica, ETL, PL/SQL, Netzetta, .NET, C#, BIRT, Java and Pearls in a corn shell and the Ruby earrings you asked for–although I can’t for the life of me understand why you want jewelry experience. Interviews start tomorrow! And all of them will take $68K so budget is good.” The hiring manager is assured that this batch of half-qualified liars were the best available people without really knowing what was said in the phone screen. The hiring manager then focuses on the core requirements in the face-to-face meeting and can’t spend too much time evaluating the nice-to-haves. When time eventually comes to use Informatica or whatever, there’s a new version so the liar gets sent for company-paid classroom training. Deceit pays off twice.

    Please stop feeding the liars. You might ask experienced BI people what they hope their candidates had actually learned from prior jobs.

    • A Current Data Architect

      I completely agree with Experienced Developer. This post actually recommends lying in an interview and appears to be written by someone with no formal Data Architecture experience. Today you can see that the majority of jobs listed for Data Architects invariably require knowledge of tool X or Y rather than on the base skills that are necessary to actually do the job. In one of my jobs, the lower managers derisively referred to it as PP&C (Pretty Pictures and Charts). The skills you need to be a good Data Architect are almost never discussed and trivial skills you can learn without any experience or ability are focused on exclusively.


        “PP&C”–love it!

        Thanks for the support. I guess I’m going to download ERWin community edition and start entering data models to show at interviews, but it seems like the challenge isn’t in documenting a single custom model (since it’s in the data dictionary) but managing interactions.

        By the time a company gets big enough for a data architect, they’ve got several “best-of-breed” systems and legacy applications. You’re integrating your HR system with network identity/security management, integrating ERP data with some “legacy” operational factory-floor system or Salesforce, feeding a data warehouse/data appliance, then querying in Microstrategy, QlikView-Spotfire, Access, Monarch or Excel/DAX. Then there’s the predictive analytics group that wants custom query rights so they can get six years of transaction data for SAS or R analysis.

        The package specialization is daunting–I have one book “Business Objects BI Security”–hundreds of pages just on managing the security of the reporting system.

        Any thoughts on how this is done in the real world?


  2. Michael Milligan

    Nice article for tactical advice on how to present your knowledge, understanding and even wisdom to the interviewer. I agree with all of the Experienced Developer’s excellent points. I’m leaving for an interview in a few minutes and, refreshingly, the requirements actually call for understanding rather than experience with a particular software package (other than Oracle). Experience with specific ETL tools is not even asked for, but rather a sound understanding of the methodologies and the circumstances in which each may be appropriate.

    I would say to Experienced Developer that, in addition to having a deep understanding of data modeling – which he clearly does have – knowing a data modeling tool such as ERwin or one of the other mainstays will add a lot. For example, how do you define and enforce your ontology in Visio? In ERwin, as well as others such as E/R Studio, PowerDesigner Data Architect, IBM Data Architect and Oracle SQL Developer Data Modeler, there are facilities for class names and naming convention enforcement.

    But I totally agree with Experienced Developer that having a deep understanding of data modeling in its various approaches, like E/R, dimensional, hybrid, etc., is crucial.