Numerify: Using Systems of Intelligence for IT Pruning

For a long time, I’ve been interested in why it is that larger companies, like Google and Facebook, that have nearly limitless resources and technology infrastructure, are all so seriously focused on pruning their technology portfolios at all levels. This focus contrasts sharply with how I believe the rest of us, businesses of more modest sizes included, approach technology, in which we act almost like hoarders, constantly accumulating new tools, while never getting rid of the old. The result is that we end up with a cluttered closet of technology that we never have the resolve or time to clean out.

I’ve written about this extensively in the past, both in a series of Forbes articles (see here and here for examples), as well as for my technology research firm, Early Adopter Research. Yet, recently, given my interest in productized analytics, I’ve realized there are products out there that can help us prune our cluttered closets of technology.

There are lots of ways for companies to gather the information needed for accurate, effective pruning, as the number of pruning tools has never been wider and more powerful. But because of the expansion of data and the large number of assets in a typical IT portfolio, when you begin a pruning process, it can quickly become challenging to bring all these components together and present it in useful way to guide the pruning process. The companies that have been most successful with pruning have a powerful system for analytics to pinpoint their most expensive data environments and how to most effectively prune their data, assets and applications.

Pruning is often focused on the most expensive assets you have, and doing just analysis of these assets can be a good idea. But as I’ve watched the development of many products that provide productized analytics, whether you adopt the product or not, you can learn a lot about pruning overall from the vendor’s approach. In order to analyze IT pruning, I’ve taken a close look at Numerify, which can provide us some lessons about how to run a reasonable program of pruning.

Numerify is identifying a system of intelligence for IT that can greatly assist companies in gaining a better grasp on their full portfolio of technology. Numerify’s system can be used as a diagnostic and ranking system for your technology: It can help create models for how everything is being used and what applications tools are most and least valuable. Based on these analytics, companies can then run a pruning process that targets their least valuable and least used assets and develop a plan for technology migration. Consequently, this systematic approach to pruning also enables companies to develop strategies for their legacy application migration.

I sat down with Gaurav Rewari, the CEO and founder of Numerify to answer the key questions one must answer when running an IT Pruning program. Here’s a summary of our conversation, with key quotes from Rewari on the topic.

Why is pruning something that is so important to large companies?

There are two main reasons — Companies need to reduce their overall spend on costs associated to basic operations and “keeping the lights on” so that more resources can be freed up for investing in innovation. And then once you have done this rebalance, you also want to ensure you’re shutting down innovative new projects as soon as it’s clear they’re not going to work so that again, those innovation dollars can be put to use elsewhere

Does IT pruning really fundamentally always boil down to app modernization and app rationalization or is it more than that?

No. It’s more than that. It also can include optimization of infrastructure and judging the effectiveness and quality of any migration (such as to the cloud). Optimizing is a form of pruning. Pruning of resources is also a constant source of focus of some companies — once a particular resource is no longer needed (applicable to both people and skills, e.g. Cobol programmers when you no longer have Cobol infrastrastructure), it’s worth shutting that down to conserve resources. Pruning of processes is also something companies should focus on — as it can also lead to savings.

What are the dimensions of pruning?

There are a range of prunings companies can conduct, including pruning of:

  • Products
  • Vendors
  • Projects
  • Skills
  • People
  • Applications
  • Code
  • Processes

It’s also important to note that consolidation is a form of pruning. You may have multiple processes, run through multiple vendor products, that achieve the same end. Reducing these and reducing duplicate work, is an effective form of pruning.

Why are large companies so committed to pruning?

It’s a cultural commitment to ensuring you have an environment where only the fittest initiatives survive. There are only so many bets you can make with innovation, so you want to be judicious with the ones you do. Additionally, individual projects can be game changers. They want to create a wide enough portfolio so that once a project proves it’s not going to work, they can shut it down and reinvest those resources elsewhere to get better results.

If you let software, applications, or other areas get too unwieldy, you end up focusing your time, resources, and attention on the maintaining that, instead of innovation. This means you’re not putting your resources where they could go to support growth.

A recent Deloitte CIO survey showed that nearly 60% of their IT budget was on sustaining operations and only 16% was spent on innovation.

Why don’t the rest of us do pruning as much as we should?

Inertia is a prime factor. Once you have an IT budget, the only incentive is to prune so that you can afford any innovation you want to support. There’s no incentive to dramatically reduce your footprint overall. People are afraid to cut.

Rewari told me: “One of the reasons they’re afraid is because they don’t have the intelligence and the data and they don’t have the framework to decide what should I cut and why should I cut it and how do I justify cutting it.”

What does a system of intelligence do to support a pruning program?

It acts as a guide to the pruning process. You gain valuable insight into what’s working and what’s not, what’s tying up resources, and where they could be better allocated. This can spur innovation and growth.

“A system of intelligence aggregates data from all relevant IT sources into a single analytical model that can be used as the basis for decision making around app rationalization, cloud migration, service cost reduction, reduction of risk in production environments, resource optimization, and all the facets and dimensions of IT pruning that companies need to care about,” said Rewari. “It also gives you insight into the initiatives that are going to be important to you tomorrow, because, if done right,  it’s really a living, breathing thing.”

Additionally, systems of intelligence allow you to make data-driven decisions, which leads to people being more willing to accept changes or cut because there’s evidence to support the decision. It becomes a fact-based conversation.

Rewari also added: “The analytical model has the advantage that it reflects your business realities, whether it’s your organizational hierarchies, whether it’s your processes, whether it’s your IT architecture — it reflects all of those dimensions so that you can then analyze in a hundred different ways at every single level within those dimensions.  So you’re not restricted to how the data shows up in its rawest format, but you are more focusing on your business processes, your organizational hierarchies, your IT towers, your IT architectures and your dollars.  So the analytical model basically translates what is otherwise raw data into a decision support system to determine what is right for your IT business.”

Finally, it also offers a single definition of truth, which helps to prevent disagreements. As Rewari told me: “We like to describe it as the three Cs.  It needs to be complete, i.e. it has to tap into all the key sort of sources of data relevant for decision making.  It needs to be conformed, which means data from different areas can be blended together easily, and it needs to be curated, which means you need to go through an exercise of paring it down to just the most important set of facts and entities and you have one and only one definition of things like mean time to resolve or business service degradation. The goal is to really give the IT organization a lingua franca to collectively make data-driven decisions that help it deliver more value to the business.”

What metrics are specifically needed for pruning?

  • App usage
  • App effectiveness
  • App cost
  • App revenue impact
  • App risk
  • But companies should also add to this list with some of their own that help to differentiate them over the long term.

To what extent does successful IT pruning come down to the following factors: measuring value, providing incentives to change, allocating time and money for change, and then accepting the disruption of the change?

The ability to manage change and communicate it effectively is the most crucial factor.

Why has IT found it so hard to measure value?

Companies focus a lot on the customer end when it comes to metrics. But when it comes to the back office side, they don’t focus as much on ROI. Going forward, that’s not going to work. Additionally, there are many aspects of keeping an app running that have no immediate ROI, but are essential to avoid a crisis down the road. Companies find it hard to measure these. Metrics that measure and quantify maintaining IT infrastructure are difficult to establish.

What are different stages in a pruning program?

  • Set your goal for the pruning
  • Establish a way to measure your progress towards that goal
  • Prioritize applications to decommission
  • Develop a compelling communication and migration plan
  • Migrate legacy business off of prioritized applications
  • Decommission legacy applications

How do you break down objections to change and app retirement related to pruning?

You have to make it so that it’s not personal. As Rewari said, “This app is not being used often, it’s got high support costs, and guess what, it does not have a high application value score.  Why are you hanging on to it?”

What are the benefits of getting this right?

Speed to innovation is the primary one. By reducing your footprint, you spend less time and resources on maintaining legacy applications, freeing you to invest in innovation and growth.

On the people side, it also leads to better job satisfaction, as no one likes to work on maintaining old systems instead of working on more innovative things.

Should pruning become part of every transformation or migration project?

Yes it should be. But you should also be pruning applications that already exist and provide value. This process should be ongoing. You want to have tight control over what makes it into an app and what doesn’t. You want to be fully aware of what an app truly costs and maintain.

What policies or practices lead to self-pruning behavior and a culture of pruning?

Having an incentive or feedback system for development teams that rewards quality of their code in production, rather than just the quantity of the code written. You also want to use data to create awareness of waste vs. effectiveness in real time.

Innovation should be incentivized within teams so that teams spend some of their time looking at their budget and deciding how to prune to free up resources that can be used to spur innovation.

What are the signs of success or failure in an IT pruning program?

The two main metrics are:

  1. Have you truly freed up resources for innovation?; and
  2. Have you truly lowered your costs on the people side?

But some other signs are business user and  developer satisfaction and a reduction in outages.