Jira spike template

One question, a fixed amount of time, and an agreed output. Without those three, a spike is open-ended work under another name.

Free to copy Standard team style

The template

Change it in the generator

Spike

Spike: [question to answer]

Priority
Jira’s default
Labels
spike

Description

Question

[The one question this spike answers.]

Timebox

[How long, for example two days. When it runs out, stop and write down what is known.]

Why it matters

[The decision or story waiting for the answer.]

Output

[What comes out: a recommendation, a prototype or a comparison. Agree it before starting.]

Done when

  • Question answered, or the reason it could not be
  • Findings written up and linked here
  • Follow-up issues created
Show it as Markdown
### Question

[The one question this spike answers.]

### Timebox

[How long, for example two days. When it runs out, stop and write down what is known.]

### Why it matters

[The decision or story waiting for the answer.]

### Output

[What comes out: a recommendation, a prototype or a comparison. Agree it before starting.]

### Done when

- [ ] Question answered, or the reason it could not be
- [ ] Findings written up and linked here
- [ ] Follow-up issues created

What a spike is, and what a good spike template contains

A spike is a timeboxed investigation. The team does it to answer a question it cannot estimate or decide without, not to ship something. A good spike template holds that one question, the timebox, the decision waiting for the answer, and the output agreed before anyone starts. A short checklist says when it is done: the question answered, the findings written up, the follow-up work created.

The summary states the question: “Spike: can search run on Postgres full-text?”, not “Search research”. The answer goes into the findings and into new work items, so a spike ends with a decision or with work the team can now estimate.

When to use it

Use it when the team cannot estimate or decide something until someone finds out more: a library, an approach, a performance question, a vendor, or what users actually do.

Jira has no Spike work type out of the box. Teams that want spikes counted on their own add one; many create spikes as tasks with a spike label instead, which is what the label in this template is for. If the work is already understood and only needs doing, it is a task or a story, not a spike.

What each section is for

Question
One question. Two questions are two spikes.
Timebox
The timebox is the point: when it runs out, the spike ends with what is known.
Why it matters
Names the decision waiting for the answer, so the answer can be judged by whether it settles it.
Output
Agreed up front, so nobody expects production code from a spike.
Approach Detailed
Useful when someone other than the author may take the spike.

Sections marked Detailed are not in the template above. Switch them on in the generator, where you can also edit the text and the checklist.

Jira spike examples

Four spikes from different teams, each written with this template. Each asks one question, sets a timebox and agrees its output up front. The Approach section is from the Detailed version; leave it out when the person doing the spike already knows how.

Software team: technical feasibility

Summary
Spike: can search run on Postgres full-text?
Question
Can product search run on Postgres full-text search with results good enough to drop the separate search service?
Timebox
Three days.
Why it matters
The search service’s contract comes up for renewal next quarter. Whether to renew waits on this answer.
Output
A recommendation, with the results of the same searches run against both, side by side.
Approach Detailed
  • Index a copy of the product catalogue in a test database
  • Run last month’s most common searches against both
  • Note every search where the results differ, and which is better
Done when
  • Question answered, or the reason it could not be
  • Findings written up and linked here
  • Follow-up issues created: a migration story, or a task to renew the contract

Performance: find the cause before sizing the fix

Summary
Spike: why is the monthly report slow for large accounts?
Question
What makes the monthly sales report slow to open for the largest accounts, and what would fix it?
Timebox
Two days.
Why it matters
Customers on the largest plan have complained. The team cannot estimate a fix until it knows the cause.
Output
The cause, with the measurements that show it, and one or two ways to fix it with a rough size for each.
Approach Detailed
  • Profile the report against a copy of a large account’s data, never the live one
  • List the slowest queries and how often each runs
Done when
  • Question answered, or the reason it could not be
  • Findings written up and linked here
  • Follow-up issues created for the chosen fix
  • Profiling code removed, or clearly marked as a prototype

Integration team: vendor and API evaluation

Summary
Spike: which SMS provider meets our delivery and data rules?
Question
Which of the two shortlisted SMS providers can deliver to every country we serve and keep message data in the EU?
Timebox
Four days.
Why it matters
Our current provider is ending the plan we use. The integration story cannot be estimated until we know which API we are writing against.
Output
A comparison of the two providers against our requirements, and a recommendation. Pricing and terms go to whoever signs the contract.
Approach Detailed
  • Read each provider’s API reference and data processing terms
  • Send test messages from each sandbox account to the team’s own phones
  • Ask legal to check where each stores message data
Done when
  • Question answered, or the reason it could not be
  • Findings written up and linked here, with legal’s view
  • Follow-up issues created for the integration

Design team: a UX research spike

Summary
Spike: why do new users stop at step three of set-up?
Question
Why do new users leave the set-up wizard at the third step?
Timebox
One week, including five short interviews.
Why it matters
The wizard is due to be redesigned next quarter. The team needs to know what to change first.
Output
A short write-up of what people said and did, the problems that came up most, and recordings for those who agreed to be recorded.
Done when
  • Question answered, or the reason it could not be
  • Findings shared with design and product, and linked here
  • Follow-up stories created for the redesign

Each spike ends by creating work: a story, a task, or a recorded decision not to go ahead. Link those work items from the spike, so the reason behind them is one click away when someone asks later.

How to make this the starting point in Jira Cloud

Jira Cloud has no issue templates as such. These are the native ways to start new spikes from this template, as Atlassian’s own pages describe them, and where each one stops. Our guide to Jira issue templates in Jira Cloud compares them in more detail.

First: a Spike work type, or a task with a label

Jira has no Spike work type out of the box. In a team-managed project, work types are set up under Space settings → Work types on a current site (Set up work types). That page says you need the Administrator role, with the Administer spaces permission, to change them. In a company-managed project, ask your Jira administrator. The other way is to create spikes as tasks and add the spike label, as this template does, so a search or a board filter can still find them.

Team-managed projects: a default description

Each work type in a team-managed project can have a default description. Atlassian’s recap on JRACLOUD-4812 gives the path: Project settings → Issue types, select Description, and write the default text. If you added a Spike work type, paste the template there. If spikes are tasks with a label, the Task default belongs to your task template, so this one has to be pasted in by hand. Either way it is the description only, the label is added by hand, and it cannot contain a table (KB).

Company-managed projects: Atlassian’s workaround

Here the system Description field cannot have a default. Atlassian’s workaround gives a custom Paragraph field the template as its default value, puts that field on the Create screen in place of Description, and copies it into Description with a post function on the workflow’s Create transition. It changes a custom field, screens and a workflow, and it gives each work type one template, so it suits spikes only if they have a work type of their own.

Cloning a reference spike

Keep one spike that is never closed, with this template in its description and the spike label set, and copy it with ••• → Clone when you need another, as Atlassian’s Clone page describes. The copy’s summary starts with “Clone”, and the reference spike turns up in searches and on boards like any other. Anyone who can edit it can change the template for everyone.

Automation after the spike is created

An Automation rule can write the template into Description: the Work item created trigger, a condition on the work type, and an Edit work item action, as in Atlassian’s KB. It runs once the spike exists, so the person in the Create dialog still starts from an empty description. Every trigger, condition and action that runs counts towards your Automation usage, and a rule on Work item created runs for every new work item, even when its condition says no (usage doc).

None of these lets someone pick the spike template in the Create dialog when spikes and tasks share a work type, or choose between a quick technical spike and a longer research one. That is what an app adds.

Want this as a template people pick in Jira?

Issue Templates & Create Defaults for Jira gives each project a Templates page. Pick a template there and Jira’s own Create dialog opens with the summary, description with its checklist, labels and priority filled in, so a spike created as a task still gets its spike label. From version 2.9.0 a template can also set the assignee to the person using it, a due date a number of days ahead, and components; for a spike, the due date can mark the end of the timebox. The person can change anything before creating, under Jira’s usual rules for required fields and permissions. Set a default template for a work type, and the Create dialog names it. The app writes nothing into a form by itself. It can also set custom fields (text, number, date, dropdowns and more), and it runs on Jira Cloud only.

To take a template from this page into the app with its formatting, paste it into an issue’s description, then, as a project or Jira administrator, choose Save as template from the issue’s ••• menu.

Questions

What is a spike in Jira?

A timeboxed investigation: work done to answer a question the team needs answered before it can estimate or decide something, such as whether an approach will work or which vendor to use. What comes out is knowledge, such as a recommendation, a prototype or a comparison, not a finished feature.

Does Jira have a Spike work type?

Not out of the box. In a team-managed project, an administrator can add one in the project settings. Many teams create spikes as tasks with a spike label instead, which is why this template includes the label.

How long should a spike be?

As long as the question deserves, agreed before starting: often a day or two, and rarely longer than a sprint. When the timebox runs out, the spike stops and records what is known, even if the question is still open.

What happens at the end of a spike?

The findings are written up and linked from the spike, and the follow-up work is created as new work items that can now be estimated. If the question is still open, the spike says why. Spending more time on it is a new decision, and a new spike, not an extension of this one.

Or build your own in the Jira issue template generator, and see what Jira Cloud does on its own.