Measuring What Matters

Measuring stick

By Maureen Vavra

Our clients often call on us at StrataFusion Group to help with Business Intelligence projects to validate and monitor major corporate initiatives. BI and Big Data have become fundamental in managing profitability and improving efficiency for business today. Managing with data can also make a big difference in smaller internal change management or projects, too.

It doesn’t need to be too complex, and the payback is clearer understanding up front and a better grasp of what an initiative needs to accomplish to be successful. I find that working with clients to identify a few key measures is a good way to quantify the level of change they want and can realistically achieve.

For projects or initiatives, define success before you start – which needle do you want to move?

  • Any big effort has to boil down to a few quantifiable outcomes, usually along the lines of the project management mantras of cost/schedule/performance. Define your outcomes and your tolerance for variance. Make it all explicitly clear.
  • For any change, the Performance area is the most important. Before you start to plan a project or a major change, define the outcome, and what success looks like. Make sure you can measure impact.
  • Keep a close eye on the business imperatives you don’t want to adversely impact – and reward people for maintaining their high standards. It’s called “managing multiple priorities.”

Measure what matters – don’t spend a fortune on reporting in the early stages

  • Ask the people doing the job or receiving the service what really matters.
  • If any effort is monitoring more than 4-5 key measures, you have overthought it.
  • On a “replace functionality” project, cut the number of reports you generate in half (even if you just did that.)
  • Think about metrics as if you’re driving a car – how many dials can you watch and still get somewhere?

Got a problem or roadblock? Value the naysayer

  • Vocal employees are saying what others are thinking. Examine the barrier: real? big? how can it be tackled? Make it a measure you knock down.
  • Accept it: good employees who are negative about a new project or change usually have a point – get them to quantify why they think something won’t succeed and help turn it around.
  • Challenge critical thinkers deliberately: to suss out what could go wrong, develop risk mitigation tactics, and help quantify and test the system for failure points.

Take a victory lap that boosts morale

  • Set solid milestones for internal initiatives, monitor and acknowledge when they are met. Recognizing specifically what worked, why something is more efficient, provides better job clarity.
  • If something fails and you catch it early, credit your measures for giving you an early warning – that’s a major value add.
  • Tie your measures to the bottom line, saving or making $, improving quality, increasing Customer Sat – it keeps the CFO happy.
  • People like to be rewarded for specifics, to know what to do to succeed – show them that something they did made a difference.

Finally, once a change management initiative or project completes, having quantifiable data can be invaluable in assessing key learnings and planning for follow-on activities as well.  If there is a larger BI or reporting effort required, the foundation has been set.

Big Data that Support Key Business Results

By Doug Harr

Word Cloud "Big Data"

CIOs have a tremendous opportunity to harness Big Data. But CIOs are also wary of buzz words and heavily marketed trends which often lead to pursuits that are secondary rather than those aligned to key results. And while it may not be clear to everyone in the executive ranks, CIOs are keenly aware that all systems (not just business systems) in an organization spew out data, much of which can be mined for useful information. When I was CIO at Splunk, we called this systems-generated data “machine data” and I had the chance to witness just how many brilliant things can be done by harnessing it. So when and where does it make sense for CIOs to embark on data driven projects? How can a CIO choose where to focus efforts?

In a typical corporation, CIOs look after everything from business applications, operations and infrastructure, security, and the infrastructure that supports their web presence. Looking across the vast portfolio of services they support, a CIO’s primary concern will be to properly implement capabilities, and then manage them in such a way that the business is effectively and efficiently supported. Taking on analytics becomes the next layer to tackle once each fundamental service is in place. Where the rubber meets the road is when you can use machine/big data to determine more than just the status of your infrastructure. That is, when you can see the opportunity to mine data for services that support the portfolio and ultimately the corporation’s key results.

Getting Started

Select a Use Case: Focus on high-value use cases first. External-customer facing use cases are particularly well suited as first forays into data mining programs. Making the customer experience as compelling as possible is key for all organizations. Developing deeper insights into this experience has enormous potential and will garner support from your marketing team and other internal customers.

Work with Your Internal Business Partners: Meet with your internal team, and departments such as marketing and engineering, to select a use case they care about. Choose a project that will impact their external customers—typically the customers of your company. While internally focused use cases for Finance, HR, Sales or other teams can be instructive, prioritize programs that address the company’s core product or service and customer experience.

Put the Technology in Place: Don’t place all your bets on one solution. Consider your approach and look at real-time products (such as Splunk), cloud offerings, and batch-oriented systems (such as Hadoop). Before you make any purchases, do a proof-of-concept. Ensure you have support staff from the vendor working with you and try a sample set of your data in their engine.

Review the Reports: Step back and review reports from the solutions you are considering. Analyze the insights, both qualitative and quantitative. For example, if you use a customer support system for your proof-of-concept, ask questions like these:

  • How long does it take a customer to get through the online sales cycle? How much time elapses from engagement to first customer support call?
  • How long are customers spending in our systems?
  • How many orders are placed per month? What’s the typical amount of time it takes to book an order? How long does it take to book an order at month end?
  • Does it appear anyone is trying to infiltrate our systems?

Demonstrate What You Can Produce: Share your proof-of-concept results with your internal team. There’s no greater fun than giving your sales and marketing customers something they didn’t have before, something that helps them make better decisions more quickly. Note that there are some use cases you will never be able to share widely. For example, security use cases can only be shared with security personnel and auditors.

Delivering Value

Bringing Big Data programs into your company is worth the effort. These data can tell you things about your business and systems you can’t learn any other way. Chosen and managed carefully, these programs can improve customer service (internal and external) provide a qualitative view into the customer experience, offer clearer insight into the products and services, and even enable a company to better understand its own employees.

Doug Harr is a partner at StrataFusion. He has more than 25 years of technology leadership experience both as an executive-level technology practitioner and in senior leadership roles for professional services organizations. Contact him at dharr@stratafusion.com; follow Doug at twitter.com/douglasharr.

More BIG Data

Not Just a Buzz Word for CIOs

Doug Harr

Big Data 2

 

What do CIOs do with Big/Machine Data?

In 2010, most of us were deleting machine log data from our systems as soon as it was clear that processes had survived the night – very frequently this data was being tossed in the trash daily. Now a short four years later, we’ve all learned that there is information in that data, and that by saving it and using search and analytics to mine it, an amazing number of things are possible.

splunk-logo

“splunking”

As CIO at Splunk (a rapidly growing company that makes a platform aiming to make machine data available, usable and valuable for everyone) the first example I saw of the use of the the solution within company itself was related to their go-to-market model. Splunk had and has a “free-mium” model where customer and prospects can download Splunk software to their PC/Mac or host, then run machine data into it to search or analyze the data. We were “splunking” those downloads – for example taking the Apache web log from the Splunk web site, contact feeds from our CRM system, Salesforce, for a lookup table, and communications back to our site which come back from Splunk itself once up and running. With just these three types of machine data records, one being a “lookup” table to enrich the data, we were able to produce an amazing array of analytics and reporting used by IT, product management, marketing, and the others in measuring the download experience, uptime, and capacity, but also the actual sales pipeline, and understanding the company’s prospects.

Downloaded Experiences – Visualized

Downloads Experience

Stats

Since IT was responsible for making sure that the free Splunk software download function was operating properly, we were interested in the download experience – things like average minutes per download, and how that differed by platform.

 

 

 

We also liked seeing activity via geo-mapping, and other dashboard visualizations, as shown below:

Downloads by CRM Region

CRM Map 2

 

 

 

 

 

 

Real-time Data – Driving Business Excellence

Over the years the use of Splunk internally was expanded to address needs for both IT and business constituents providing customer insight, protecting against intrusion and malware, enhancing operations effectiveness, and other uses, falling into these categories:

  • Monitor and manage infrastructure – capacity, uptime, project delivery
  • Deliver application management – health of business apps, usage statistics, even some missing reporting
  • Provide analytics on security posture – identify and eradicate malware, APT’s (advanced persistent threats), and other threats
  • Provide business analytics – most of these derived by departments – people in sales, marketing, and engineering analyzing business trends, product delivery, customer support and more
  • Internet of Things – we even “splunked” our headquarters building to review temperature and C02 levels

These examples roughly match the broad spectrum of what can be done when ingesting and analyzing machine data in real time. Stay tuned for more examples in posts to come. Now with StrataFusion, I will be consulting and teaching more on these topics!

 

 

Big Data

Not Just a Buzz Word for CIOs

Doug Harr

Big Data box image

Four wonderful years at Splunk as CIO. Splunk? Splunk is a rapidly growing company that makes a platform aiming to make machine data available, usable and valuable for everyone. While there, I built the IT and Real Estate/Facilities teams and solidified an “all cloud” business applications portfolio. This advanced my knowledge of all things cloud, this time including the appropriate use of Amazon’s EC2 (Amazon Elastic Compute Cloud*) for compute and storage needs. At Splunk everything but Engineering applications were delivered via cloud subscriptions, and half of the compute and storage needed for Engineering, from EC2. More on that in future posts.

Harness Opportunity

The most impactful thing I learned at Splunk is the tremendous opportunity CIO’s have to harness what the market is calling “Big Data” and which Splunk refers to as “machine data.” In this context, “machine data” can be thought of as system logs, sensor readings, results of polling and measuring machine behavior. Every computer system, storage, device, web, app, and database spews forth machine data – much of it delivered via a constant, real-time stream from the machine – and almost all of it in text format. The original application of Splunk was for data center management. What was built worked equally as well for application management, security, business and web analytics, and more recently, to monitor and analyze devices connected as “the internet of things.” Results come from searching through the data and formulating analytics from its content – ranging from things like “are the machines up? Are there signs of imminent failure? Are there attempts to infiltrate and hack the system? …..to “Has Joe taken his heart monitor off?”   Uses are limited only by the imagination. What can you do with you data? Learn more in my next post.  Or, visit me at StrataFusion.

*Amazon Elastic Compute Cloud (Amazon EC2) provides scalable computing capacity in the Amazon Web Services (AWS) cloud.