Planning and managing expectations

Learning from other projects can help you to plan your standards development project and manage expectations in the community.

All projects that create and develop standards are different, but most will have a similar shape. We have found that project teams can easily underestimate the time it can take to develop a standard.

Standards projects are not always straightforward. Some important aspects – like getting people to agree on or adopt the standard – are hard to control because they rely on decisions made by others.

The steps here offer a guide to help you plan roughly how long you should expect each of the phases involved in developing a standard to take.

Scoping and starting

How long does it take? Between a few weeks and a couple of years.

Why does it take so long? Although the tasks involved in the scoping phase sound like they can be completed fairly quickly, this depends on the choices, requirements, stakeholders etc being clearly set out, which takes time. In some cases, like in the Brownfield Site Register Open Data Standard, this phase will include a (potentially lengthy) process of setting a strategy. In other cases, pilot programmes can help to work out scoping-stage requirements in terms of features, budget, schedule and resources.

Could it go faster? Yes, depending on the type of project, and maturity or experience of the sector or group that is working on the standard. For example, if you are iterating on an existing standard, most of what is needed at the scoping stage is already available.

Development

How long does it take? Almost always a year or more – complex standards can take several years to be considered complete. Most of the examples in our case studies have had a development phase (ie deciding on and writing the technical specification for the standard) of one to two years.

Why does it take so long? The task of developing the standard – ie deciding on and writing the technical specification for it – is the easy part: the initial data schema for the General Transit Feed Specification (GTFS) standard was generated in approximately one hour. Everything else is about agreement and ensuring quality, and those things take time.

Could it go faster? It could, but beware. Rushing decision making or cutting corners on testing is likely to result in a standard that is unfit for purpose, not adopted or badly implemented.

Launch and adoption

How long does it take? While the launch itself will take a set time period, adoption can be an ongoing process the standard being adopted over time.

Why does it take so long? This depends on the terms of the adoption, the number of people or groups expected to adopt the standards, and whether they have been involved throughout the development process. For Open Banking, a managed rollout was organised between January and April 2018. Other standards get adopted over many years, as the ecosystem around them evolves.

Could it go faster? This is typically faster for mandated standards, where the adoption is a requirement for a service licence condition, than for voluntary standards, where adoption will be more organic.

Review, update and retire

How long does it take? Could be days, could be years.

Why does it take so long? In some cases, it is important to provide stability, so that implementers and others can build on the newly released standard. And while iterating on a standard can be faster than creating a new one, it still takes time.

Could it go faster? In the case of “living standards” (a specific type of always-evolving standard), the process of reviewing and updating can be very fast. The GTFS & GTFS-RT case study shows how changes can be requested and adopted within days, so long as there is evidence that they are widely useful, tested and approved by the community.

How to use this guide

There are a number of ways for you to learn more about the creation, development and adoption of open standards for data.

About this guide

This guidebook helps people and organisations create, develop and adopt open standards for data. It supports a variety of users, including policy leads, domain experts and technologists.

Something missing? Suggest a change or addition