Decision Silos

Removing Silos in Planning - Part 5

In previous parts, we described process silos,?data silos, and information silos. In this article, I will describe the last of the 4 technical types of functional silos in organizations. As you may have noticed each next silo type got increasingly complex compared to the prior one. This pattern continues with this part. Decision silos are the most difficult to overcome.

Removal of all the prior silo types is a prerequisite to removing decision silos. When decision silos have been removed all parts of the organization work towards the same overarching goals. Consistently, top to bottom, left to right. But if all the parts of the company are not basing decisions on the same timing, same version of the truth, with an adequate understanding of their degrees of freedom, how can they possibly come to a result that is best overall? If you remove decision silos before removing the other silos, you align on the "what" without knowing the "how". It is better than not aligning at all, but not nearly as impactful.

Misalignment of decisions can be found all across an organization (and like the other silo types, also between organizations). When production planning maximizes batch sizes for manufacturing efficiency they bloat inventory, directly competing with inventory management's objectives. When finance reduces the budget for inventory they impact customer service levels, competing with service and sales objectives, and as result they see their own numbers deteriorate due to loss of revenue. Through lack of decision alignment, the organization is fighting itself instead of moving together. All the conflicting local optima sum up to something far from a global optimum.

A great article on decision silos appeared in Harvard Business Review back in 2006: Who Has the D?: How Clear Decision Roles Enhance Organizational Performance. Definitely worth a read. It covers these silos from a functional perspective. In this series, we take a different, decidedly technical, angle on silos. In my opinion, technology needs to be a pure enabler of business, and silos cannot be properly removed without proper technology enabling it. I consider the content discussed here as foundational to achieve the functional goals.

Decision silo removal requires alignment across at least 4 technical concepts:

• Metrics - What do we measure, and how.
• Objectives - What value of our chosen metrics do we target, what are the penalties of deviating from those, and what boundaries are we not allowed to cross.
• Tasks - Which processes does the organization perform to operate and achieve its objectives.
• Outcomes - What combination of objectives best serve our desired business impact.

Outcomes are the ultimate purpose all of the above serves. An outcome may be "delighted, retained, and profitable customers", or "net negative carbon footprint". They are typically "fuzzy" broad visionary concepts. They impact all or most parts of the organization. They are executed through tasks performed within the organization and made concrete by choosing a collection of objectives that would achieve the outcome. The various objectives could be assigned different levels of contribution to the outcome, such as through balanced scorecards. some objectives will be served by only one process, but most will be distributed across multiple processes or departments. Here's a simplified example:

Outcome: delighted, retained, and profitable customers

• Task: Order entry fast and without error. Objective: 99.9% flawless entry.
• Task: Orders completed in full and on time. Objective: 99% OTIF.
• Task: Competitive and differentiated product. Objective: 4.5+ star rating
• Task: Competitive and differentiated pricing. Objective: <10% closing failure due to price
• Task: Invoicing and collection timely and error-less. Objective: <30 days order-to-cash.
• Task: Relationship valued and managed. Objective: <1% annual customer attrition

Naturally the above is much simplified. One task will generally contribute to multiple objectives, and one objective will generally be served by multiple tasks. And all will need hierarchical relationships. Tasks, sub-tasks, sub-sub-tasks. Objectives, sub-objectives, proxy-objectives.

This brings us to metrics. Objectives operate on metrics. For example, OTIF (On-Time, In-Full percentage) is a metric, but a target of 99% OTIF is an objective. You cannot quantify objectives without metrics, and you cannot objectively quantify outcomes without objectives. To achieve an outcome you need to start with getting the metrics right.

The problem is that most metrics cannot be measured universally across an organization. Manufacturing capacity utilization has no meaning in invoicing. Customer service levels have no meaning in raw material shipping. This is one reason why companies have a hard time aligning their processes.

In terms of metrics, departments speak different languages.

The only universal metric most companies believe they have found is cost. But cost is a horrible metric. It only covers one side of the financial balance coin, which in turn is only one of multiple coins that contribute to success. To minimize cost, the optimal solution is to close up shop. And cost is not the same thing for everyone. For some it is variable, for others it is fixed, for yet others both, or part of both. For some cost is purchase cost, for others handling costs, or holding costs. or... the list goes on. Cost is not universal, and cost is inherently conflicting.

So what metrics should you use?

Well, it depends. Most assuredly, there will be many. Each process will need to be measured by multiple complementary metrics, which will more often than not, also be conflicting. All metrics must be aligned horizontally across the organization and vertically up and down the organization. At least the ones used in objectives. This is not a trivial exercise, and cannot be copied from other companies. One type of metric that can be used universally is flow, which has many appearances in different areas, such as throughput in manufacturing and cash flow in finance. This metric can serve as a base to build upon with other complementary metrics.

Once the metrics are decided, the targets, penalties, and constraints on objectives can be set. Where metrics need to ensure consistency across the organization, objectives need to then set the goalposts for each process. The objectives can change over time as business strategy or tactics change. The metrics generally do not.

Whilst the metrics need to be chosen objectively, objectives are subjective

Head spinning sentence! Whatever values are set on the objectives then must be aligned throughout the organization, and adjusted throughout whenever the values are changed. They must drive both human decisions, augmented decisions, and automated decisions. All in the same direction.

Note that there may be many additional metrics that are not part of shared objectives. For example, whilst resource utilization may not be an objective for manufacturing, it will still need to be monitored. When it is consistently very low there may be potential for cost reduction. When it gets too high, throughput will start to suffer. When it fluctuates too much, laborers' schedules become unmanageable. This will drive important decisions within manufacturing even if it does not contribute directly to the overall strategy. This alleviates the need for cross-functional alignment but also prohibits placing any objectives on the metric since those would break the overall alignment.

To summarize it all. Outcomes express the strategy. Tasks simply are. They are the things we do to run the business. They are mapped to objectives they contribute to. Together they roll up into the outcomes they serve. And metrics are the foundation upon all of which this is built. Most companies have done all the above. Just on the fly, gradually over time, leading to decision silos. To remove those silos it needs to be done deliberately and coherently. Further into this series, I will present a framework that can be used as a guide.

This article is part of a series covering the 4 types of functional silos in planning:

• Part 1 - Introduction
• Part 2 - Process Silos
• Part 3 - Data Silos
• Part 4 - What are Information Silos
• Part 5 - What are Decision Silos (this article)
• Part 6 - How To Integrate Information
• Part 7 - A Decision Framework Blueprint (link to be added once published)
• Part 8 - The Data-Driven Enterprise (link to be added once published)

Find all my articles by category here. Also listing outstanding articles by other authors.