Ad-hoc reports Anonymous
A Five Step Recovery Program
Ad-hoc Report requests are the bane of most analytic shops around the globe. They tie up significant resources, up-end prioritization, and leave most career analysts underwhelmed with their job description. Worse still they form a negative feedback loop where each request delays the creation of more permanent solutions. The absence of those solutions, in turn, lead to more requests.
To undo this death spiral of inefficiency and dissatisfaction requires awareness, action, and persistence. Understanding the five levels of reporting and analysis and how best to resources them will help you develop a strategy to overcome it.
Queries
Any company that invests in a database should also be investing in asking it questions. Those questions, in data parlance, are known as queries. They are not typically unique to analytic shops. Developers, engineers, and database administrators are likely running numerous queries, too.
Queries are typically run via Structure Query Language, commonly known as SQL (pronounced Sequel). Recently, however, there has been a trend toward non-SQL databases which use different means to query them.
Queries are static, point in time, events. They are run most often with the goal of providing counts, sums, and averages to fairly simple questions. But queries are not so simple. Miscode a query and the answer you receive will range from ridiculous to horribly misleading. Note- complex queries are the heart of most reporting and business intelligence, but we will cover those later.
To optimize your team/companies use of queries, consider the following. Leave queries to junior and entry-level analysts with senior oversight. This should keep them cost effective while providing training opportunities for the team. Second, any query that you need to run more than twice should probably be in a report.
Reports
Reports are typically composed of a series of static queries. They are not interactive. Interactive reports are just a bad name for business intelligence (more later). They should also be a scheduled and repetitive process. This does not mean you can’t iterate, but the unscheduled, ever changing report is the very poster child everything this article is trying to address.
Another common myth is that almost anyone in your company can create a report. This is a mistake! I will write more about this in a later article, but consider the following questions. Would you allow your physician to conduct your heart surgery? Do you want your urologist prescribing you depression medication? Would you hire the local handyman to build your house?
Reporting can be optimized one of two ways. You can give it to the analytic team or you can give it to a business intelligence team. But you must be sure that reports are overseen by experienced analysts who have specialized in reporting. Whatever you do- do not leave it to the business, the IT department, or this summer’s interns!
Decks
Decks can and should be a tool for any analyst team. Most decks should get support from analyst and reports. But decks do not need to be built by your analysts!
Decks tell stories, hopefully using well-developed data. Storytelling may be a strength of some of your analysts, but it should also be a strength of many of your marketers and program managers. Unlike reporting, businesses are less inclined to believe anyone can build a good deck. But often they direct their ask to the wrong team…
Business Intelligence Solutions
Business Intelligence is a complicated field. Great Business Intelligence is not just interactive reporting, although that is a component. BI Developers are expert analysts in their own right and their solutions represent some of your most important IP.
A great BI solution is a model of your business. It allows you to gain insights, track change, and forecast future outcomes. BI should include segmentation and benchmarks. It includes forecasting, scoring, and visualization. A good BI developer/designer needs broad knowledge and experience.
Business Intelligence solutions are one of the most cost-effective investments your company can make. But this is only true if they are designed and maintained by experts. Your company needs to have dedicated resources to do this right. Whether to outsource or resource internally is another question entirely.
Analysis
If you address these first four stages appropriately, what you are left with should be deep and meaningful analysis. Some of it will still be ad-hoc. Depending on your business, this may require further delineation. Disciplines like statistical modeling or web analytics can be quite varied.
It will take another article to dig into the divisions of labor and ownership for each type of analysis that you may need to support your business. But with the demon of ad-hoc reporting at bay, you are now much more capable of addressing this new level of challenges. You are likely to find your P&L more capable as well.
For more information about the author see his profile on LinkedIN — George Earl
