Blogs and media are full of articles about the Internet of Things (#IoT), huge data sets (“Big Data”) , and myriad sensors. Companies are collecting terabytes and petabytes of Big Data; but, unless you sell disk drives, computers, or sensors, there’s a huge cost and not much value in simply collecting data.
The real value lies in improving the execution of your business– better customer service, increased revenues, and higher margins. The trick is to define organizational habits as a forcing function that will leverage the new technology to drive the desired outcome, and not get lost in the process.
Over my career, I have observed some best-practice “habits” of organizations that are successful in getting business results from technology.
- Defining the business problem(s) being addressed. As with any business initiative, the clearer the goals, the better the results. Companies with the most success using retailer demand data (POS, INV) have clear corporate objectives around out of stock, inventory turns, margins, etc. These goals are defined down to the individual team level, measured weekly or monthly, and reviewed at least quarterly.
- Aligning the organization to solve that business problem. Having the right organizational alignment—helps drive the desired behavior. The team reports to an executive sponsor, at least for that set of results or metrics. A program manager ensures that meetings are structured, regularly scheduled, and productive, tracks progress toward the objective, and ensures a regular reporting cadence for the team and keeps the team abreast of changes.
- For new problems or significant changes in the problem scope, they run a limited pilot. A pilot of new technology allows the organization to determine if it is fit for purpose, both for the business objectives and with respect to key organizational processes, such as security. A new technology that bypasses security requirements would be an organizational disaster, as some executives have recently learned. A pilot will allow you to both test the technology and discover these organizational requirements before you deploy.
- Pareto analysis of the data collection, relative to the business problem. Not all data are equal: a critical few of the metrics really matter to any project. Making sure that these are collected properly and understood is far more important than collecting all possible data. Helping teams understand this and focus on only the data relevant to desired results is a critical challenge for a project leader. As the organization’s maturity increases, more data can be used.
- A process for maintaining the quality of the Master and Fact Data. For both Fact and Master data, include your key business users in the upfront definition process—do not delegate it to a programmer or vendor. Each decision about naming master data attributes, the values of those attributes, and how the data are aggregated incorporates critical assumptions. Discovering these after the fact can hamper a project dramatically.
Aligning the team(s) to a common framework is critical. Give thought to
- Calendars (same calendar or mapping between calendars?)
- Standard naming conventions (“Red”, vs “RED”, vs “Rd”)
- How updates/changes are handled.
This is a lot of work and requires continuous team alignment (see point 2) to be successful, but absolutely necessary if your teams work across organizational divisions and with customers and suppliers. This work can be drudgery so make it fun – having accurate Master Data will dramatically increase your project ROI.
Make sure your fact data collection processes are robust: track meta data changes over time. Even the way the data are keyed can be important. Vendors can help you understand what the data represent; they can also help you track quality. Understanding the data in detail allows you to distinguish between a one-time event, a data quality issue, and a substantive change in the business. In a recent example, a client was tracking a new product introduction last winter. While it went well, localized snowstorms stopped sales for a few days in some stores. Understanding the data allowed the team to filter out extraneous data and correctly measure how the rollout was going. Other less aware customers might have simply noticed a reduction in sales and attributed it to a poorly performing product launch.
- An ongoing review process of the business problem and supporting data. Teams that use a monthly scorecard for tracking issues and results do far better than teams that do not. Holding teams accountable to a standard monthly measurement delivers much higher results. Scorecards enable you to intervene on issues earlier and be confident about the results.
- Pushing for results and celebrating success. Technology projects are rarely easy—each win should be acknowledge and celebrated. Team members need the encouragement they get from knowing that their projects are important and make a difference to the business.
