A successful data and analytics strategy relies on a number of moving parts; the sponsorship, culture and skills, architecture, and many more. Synergising these elements takes coherent and deliberate organisation.
In this article, we’re going to look at how data and analytics can be structured within a business, the pros and cons of different approaches, and the impact these approaches can have on the value delivered.
Data & analytics organisation at an small and medium-sized organisation (SMB)
Before we discuss the different approaches to organising data and analytics, lets first outline an overall structure commonly seen in SMBs.
For a typical SMB, there will likely be a head office with smaller satellite offices, and likely a be a number of product divisions These may be made up of different organisations brought together as part of a buy and build strategy – with separate sales and marketing teams. Lastly, there is likely to be a centralised, shared server centre that covers HR, Finance, IT etc.
Given this structure, in an SMB data and analytics can be thought of as having three organisational levels:
It’s very common for data and analytics to gain an initial foothold in an SMB via proactive individual sponsors or teams who narrowly utilise a tool or platform for a single use-case.
Teams purchase technology and upskill without any holistic view of the organisation and, over time, this leads to data silos, multiple platforms, different key performance indicators (KPIs), a lack of standard practices, and conflicting datasets.
Sooner or later, there is a ‘tipping point’ that pushes the business to pull the different strands together, adopt a cohesive approach to organising their data and analytics initiatives and maximise the value being delivered.
Finding the right approach
The following models offer three distinct approaches to how SMBs can organise their initiatives. For all of these models, there will be a central person who is responsible for data and analytics within the organisation.
In a distributed model, data and analytics capabilities are spread throughout the business, with various departments each developing a team member that has data and analytics expertise. While these experts are distributed into physically separated departments, they will typically have regular communication with each other and will adopt standardised practices.
A benefit of this approach is a low barrier to entry: team members have existing business and departmental knowledge, and existing connections between colleagues. This approach can also make it easier to keep the data and analytics function focused on priorities within the scope of the department – however, this departmental focus can also result in outcomes that are not entirely aligned with corporate aims and objectives.
If not managed effectively, a distributed model can lead to silos of knowledge and skills, and uneven adoption of standard processes.
In a centralised model, a single data and analytics team offers end-to-end capabilities for other teams within the business. This top-down approach offers a direct connection between senior management and the data and analytics team, facilitating a close alignment to business strategy and easy implementation of standard processes.
Some considerations include this approach having a higher barrier to entry, requiring new recruiting or new skills training – or, if existing resources are moved from other teams into the centralised team this can be disruptive and lead to a change resistant culture.
Over the longer term, prioritising incoming requests for the data and analytics team can become a significant overhead and a centralised approach will require careful management of competing demands.
Finally, a hybrid-federated model, offers the best of both words by combining a centralised data and analytics team with some members of this team distributed across various business functions. These distributed team members typically report to the centralised data and anlytics team lead or manager, but also need to be aligned to the management of the business function in question.
This approach reduces the barriers to entry that are seen in the centralised model. The centralised team can be scaled up more slowly while, simultaneously, the distributed team members are making impacts within individual business functions. This approach also makes it easier to establish a single source of the truth, and to establish and enforce policies around governance, privacy, processes etc.
However, the hybrid federated model also requires careful oversight to ensure that all members of the data and analytics team are comfortable and competent working both within the central team and distributed to different business functions. It also requires robust communication to ensure synchronicity between business functions and the centralised team.
In summary, while each of these approaches may suit a particular business environment, a hybrid-federated model combines the best characteristics of both the centralised and decentralised approaches.
It offers a balance between the stability, security and business alignment offered by a centralised team, and the agile development and innovation that stems from having ‘on the ground’ analytics experts distributed within business functions.
This article is part of our Data-Driven SMB series. For more information, advice and resources on how to accelerate your organisation’s data and analytics maturity, click here or contact us today.
Bill is a Business Development Director at TrueCue and is responsible for helping existing and future clients gain value from their data and analytics investments.
With over 20 years of experience in Management Consulting (AT Kearney), Industry (BOC Gases; Eli Lilly) and Private Equity (LDC; Easynet) Bill has worked in, and successfully led, teams across several operational functions including marketing, sales, delivery, project management, client success, and support.