Organisational Network Analysis (ONA)
Organisational Network Analysis (ONA) is a method used to visualise and analyse the relationships, interactions, and flows of communication within an organisation, between individuals or teams. The main goal of ONA is to uncover the informal networks that exist within an organisation. Informal networks are often not visible through traditional organisational charts or structures. By mapping these networks, organisations can gain insights into how information travels, how collaborative networks are formed, and how influential or central certain individuals or groups are within the organisation.
Historically, ONA has been seen as too complex to be practically useful in understanding and transforming an organisation. Fighting complexity with complexity just leads to confused stakeholders and poor or non-existent outcomes.
We have designed a methodology of measuring network quality by combining the simplicity of employee engagement with the power of social network science. In addition, we make this accessible with a simple system KPI, the Net Collaboration Score (NCS).
NCS is calculated in the same way as the Net Promoter Score which is a tried and tested industry metric. The NCS is a measure of every connection inside an organisation and classifies them as Positive, Neutral or Negative. The calculation is the % of Positive connections less the % of Negative, ignoring the Neutral connection. This gives you a simple aggregated metric for the quality of collaboration and allows for detailed identification of areas needing improvement in the system. Once you have identified where improvements are made you can zoom in to any part of the organisation to see the quality of the system in more detail, both internal (“My Team”) and external (Cross-team) views are made available.
This is a new layer on traditional ONA solutions which generally measure a binary system of whether a connection exists or does not exist.
Types of ONA
There are two ways of collecting network data inside an organisation: active and passive data.
Passive data, sometimes referred to as ‘digital exhaust’, collects data from existing systems like Slack or Microsoft, and is used to infer a network. While this is a non-invasive approach there are two main downsides: the level of trust in the data is low, with individuals making statements like “we use emails or slack differently so this isn’t relevant for us”; and the depth of data is limited, with no easy way of extracting the types and quality of collaboration without use of even more invasive techniques such as email or message reading which are generally considered to be over-reaching.
In contract Cophi uses Active data collection to overcome these limitations. This requires an explicit collection of data from an organisation and is generally done as a survey in the same way that you might collect engagement data. The main benefits are: completeness and consistency of data; trust in the output from key stakeholders; and a depth of data that cannot be easily found in passive analysis. Such depth of data enables focused insights on the quality of collaboration, which are essential for designing effective interventions.
Underlying behaviours and network types
As well as seeing the overall quality of an organisational network, we further explore 3 distinct activities inside an organisational network: operational, social and strategic relationships.
Operational Networks are relationships that help you ‘do’ your work.
Which people or teams do you get information from, who can help you secure faster progress on an activity. Generally this improves speed of delivery and agility.
Social networks are the driver or energy behind an organisational or team system. This covers aspects of a network like trust and other informal relationships, including psychological safety networks and feedback systems. These are the intrinsic motivators for ‘doing’ and underlying reasons for building a long term ‘vision’.
Strategic Networks inside a system are related to ‘doing in the future’ where do we want to go, how do we build a plan to get there and why are we going to have a competitive advantage. Most of this relies on a system where you can easily brainstorm new ideas, solve complex problems and enable a culture of learning.
The team at Cophi has been working on the networks that drive performance for more than 10 years. Using a process of factor analysis has refined this down to a few key questions to deliver the insight you need for effective behavioural insights and change.
To further understand the impact of these networks, we draw upon the work of thought leaders in the field, including Ibarra and Hunter, and Krackhardt and Hanson.
Ibarra and Hunter who describe collaborative network types as follows:
Operational networking is geared toward doing one’s assigned tasks more effectively. It involves cultivating stronger relationships with colleagues whose membership in the network is clear; their roles define them as stakeholders.
Personal networking engages kindred spirits from outside an organisation in an individual’s efforts to learn and find opportunities for personal advancement.
Strategic networking puts the tools of networking in the service of business goals. At this level, a manager creates the kind of network that will help uncover and capitalise on new opportunities for the company. The ability to move to this level of networking turns out to be a key test of leadership.
Krackhardt and Hanson also think in a similar framework with 3 key networks:
Communication: “Employees who talk about work related matters on a regular basis”
Trust: “Employees who share delicate political information and back each other in a crisis”
Advice: “Employees who depend on each other to solve problems and provide technical information”
Cophi’s use of Artificial intelligence (AI)
Cophi uses a combination of traditional and generative AI technologies.
Our generative AI is used to take very large volumes of text feedback from all parts of the organisation and to generate intelligent summaries and recommendations from this data.
Our traditional AI is a complex algorithm that predicts expected networks based on a number of attributes linked to the organisational hierarchy and multiple other profile and relationship based data points.
Networks are a classic opportunity for traditional AI to be applied owing to the potential for network complexity.
We use our AI to run a prediction for every possible network connection and deliver weighted benchmarks based on (as of December 2023) over 500k network connections.
Networks contain extremely high volumes of potential connections. For example, the maximum number of connections in a system is 2n(n-1)/2, where n is the number of nodes (or people) in the network. So for even a simple organisation of 100 people there are 9,900 possible total connections.
Take that to the world of predictive networks. You have to consider all the potential variations in that network, i.e. not everyone will be connected, but what is the likelihood of a connection? For this we have to consider all possible outcomes. So how many possible outcomes are there in a network of 100 people?!
For directed graphs (without self-loops), each pair of nodes can have up to 2 directed edges (one in each direction), so the formula is 2^(n(n−1)).
For a small team of 5 people the total possible connected variations is more than a million!
And the complexity grows exponentially such that a team of 100 people has more possible connections than the number of atoms in the visible universe…
So when we calculate predicted networks, we need to have an AI to handle the probabilities. This is where our AI, equipped with advanced algorithms and computational power, steps in. By leveraging the latest in machine learning and network analysis techniques, it can sift through these vast datasets to identify patterns, predict potential connections, and offer insights with unprecedented accuracy. This capability allows us to transform what could be an overwhelming amount of data into manageable, meaningful information. With this approach, Cophi enables organizations to understand their internal networks in a way that was previously unimaginable, facilitating strategic decision-making and fostering a more connected and productive organizational culture.
Summary
Organisational networks are complex and serve a wide range of use cases, including culture change and transformation, diversity and inclusion, learning and development, post-merger integration, supply chain optimisation, systems measurement, operating efficiency optimisation, and many more.
With each of these projects, the variety of networks and their applications is extensive. Some networks are naturally stronger than others, and a well-functioning network might not be highly interconnected if that serves its purpose. It's possible for individuals to be both over-connected and under-connected.
The scientific goal for Cophi is to provide the right data and insights to every individual, enabling them to generate and learn insights that are relevant to them at the appropriate time.
The quality of an organisational network can vary from the top to the bottom and across all functions. Network patterns that are suitable for one organisation may not be for another. Similarly, some networks may support short-term goals but fail to prepare the organisation for a competitive advantage in the medium to long term.
Each organisation is unique and context-specific. Only with a deep understanding of the data will you be able to form an informed opinion, back it up with data, and drive change through your organisational or team system.
Resources
How Leaders Create and Use Networks by Herminia Ibarra and Mark Lee Hunter How Leaders Create and Use Networks
David Krackhardt and Jeffrey Hanson who offer up specific questions for formal and informal ‘personal’ networks relating to Trust. Informal Networks: The Company Behind the Chart.
Rob Cross also studies the critical impact of informal social networks on organisational quality: The Hidden Power of Social Networks: Understanding How Work Really Gets Done in Organizations
