Planning for unplanned work
Intro
There are two varieties of work.
Deliberate work and unplanned work.
Deliberate work are duties and actions which have been organized and scheduled upfront. Any such work sometimes occurs in initiatives that tie into roadmaps and bigger firm targets. As a result of the work has been deliberate earlier than, you recognize what you are making an attempt to do, who’s going to work on it, and when they’ll get began.
Unplanned work, then again, are points and emergencies that seem instantly and seemingly randomly within the type of a bug report, an essential characteristic request, or an outage.
Unplanned work occurs unexpectedly, but it surely’s not surprising. You know that there shall be bug experiences, you simply don’t know when, the place, and in what form they’ll come up. The one factor that’s sure is that they’ll seem.
There isn’t any approach to escape these advert hoc forces of the product constructing course of. However there are methods to plan for unplanned work and instruments to systematically handle it. What these are and the way they work is what we’ll cowl within the subsequent few chapters.
Let’s start.
The challenges of unplanned work
Earlier than we bounce into the specifics, we must always take a step again and take into consideration what makes unplanned work so troublesome to handle within the first place.
Unplanned work has two broad phases: There’s an consumption section the place unplanned work will get reported. And there’s an execution section during which the reported issues are literally resolved. Each of those phases have their very own distinctive challenges.
Consumption is difficult as a result of unplanned work seems in so many alternative and disconnected locations. Take bugs for instance. They may flip up throughout many alternative channels, each internally and externally: social media, e mail, Slack, a gathering, your buyer help instrument, and half a dozen different locations you didn’t even take into consideration. Is each recipient throughout these numerous contact factors capable of rapidly file a bug report? Do they know the place to file it, methods to write it, and who to ship it to?
Execution is difficult on account of unsure tasks and priorities. As a result of unplanned work is unplanned, it’s not at all times clear who ought to work on it. In reality, it won’t even be clear if anybody ought to work on it in any respect. Some points are emergencies, however others will be ignored. Some requests appear irrelevant, however what if they arrive out of your most essential buyer? How do you guarantee an important work truly will get completed rapidly?
Let’s take a better take a look at every of those phases, beginning with consumption.
Phase 1: Zero-friction intake
The consumption section is all about turning reported bugs and different requests into express, actionable firm data by feeding them into your difficulty monitoring system. As a result of every bit of friction will increase the probability that one thing doesn’t get reported, it’s crucial that this course of is as easy and clean as potential.
Why is that this level so essential?
Consider your difficulty monitoring system as a hive thoughts or a collective mind that comprises essential data in regards to the state of your product constructing course of. It’s the supply of reality upon which you and your organization make choices and coordinate work.
If the info in your hive thoughts is incomplete or incorrect, nonetheless, it stops being a supply of reality. It turns into a supply of half-truths. You possibly can’t know what persons are pondering if it is too onerous for them to speak with you, and you’ll’t act on what they are saying if you happen to do not hold monitor of it.
The main target of the consumption course of on the left aspect of the Venn diagram: We need to decrease the quantity of unreported points in order that the data in your difficulty monitoring system precisely displays the truth you’re in.
So how can we get there?
Step one is to permit consumption straight on the supply the place unplanned work first seems. It’s like instrumenting your group to get full situational consciousness of every part that’s taking place. The complexity right here is that there are various totally different sources and also you need to cowl as lots of them as potential.
This is the reason at Linear we put such a robust emphasis on integrations. From the beginning, we developed integrations for the most typical sources of unplanned work. They embrace buyer help instruments like Intercom, efficiency monitoring platforms resembling Sentry, and common communication providers like Slack.
Slack performs a very essential position on this stack as a result of it’s the place the place most unplanned work surfaces. It acts as a kind of catch-all for the long-tail of unplanned experiences and requests. As others have described it, Slack is like “911 for no matter isn’t potential natively in an organization’s productiveness apps”.
The 911 analogy is attention-grabbing however not solely correct. Once you name 911, somebody will reply and take motion. Once you submit on Slack, somebody would possibly reply and take motion. You don’t know if anybody has seen your name for assist or when you’ve got simply been shouting into the void.
To enhance this space of unplanned work, we just lately launched a brand new characteristic referred to as Linear Asks. With Asks, anybody in a Slack workspace can rapidly flip a dialog right into a request that mechanically will get routed to the related group in Linear.
The core concept behind Asks is to make unplanned work express and actionable — however in probably the most frictionless method potential (keep in mind, we need to maximize the likelihood that consumption truly occurs):
Most significantly, the particular person making the request has full peace of thoughts that their request shall be seen and brought care of. Whereas we don’t just like the comparability an excessive amount of, there are loads of conceptual similarities between Asks and an emergency hotline. It’s the lacking piece that truly turns Slack into 911 for unplanned work.
Now that now we have the consumption course of coated, let’s transfer on to the execution section.
Phase 2: Efficient execution with Triage
Most difficulty monitoring techniques have frameworks for deliberate work, however no native conceptual containers for unplanned work. Whereas deliberate work is neatly organized in initiatives and roadmaps, there isn’t a clear, predefined path for what ought to occur with unplanned work. It both continuously disrupts the group’s present work or finally ends up being dumped onto an infinite backlog by no means to be checked out once more (or each).
To handle unplanned work systematically moderately than simply regionally, Linear has an essential idea referred to as Triage.
Triage is a shared inbox in your group that centralizes all incoming unplanned work. All of the bug experiences, characteristic concepts, alerts, and different requests that get reported and filed within the numerous totally different channels we talked about within the earlier chapter are added to this queue.
Triage supplies an opportunity to assessment, replace, and prioritize every difficulty earlier than it will get assigned and added to the group’s workflow. This helps to filter out irrelevant requests and duplicates.
Let’s return to our Venn diagram from earlier to know why that is essential.
Whereas the consumption course of is designed to scale back the variety of unreported points on the left aspect of the Venn diagram, the primary a part of execution section is there to confirm that no matter we’re feeding into the problem monitoring system is definitely correct.
By reviewing all points, we will filter out invalid requests, ask for extra data if essential, and merge duplicates which have been reported earlier than. This fashion we ensures that the data in your “firm hive thoughts” is not only a whole but in addition an correct reflection of actuality.
The assessment course of additionally helps to make clear what a part of the unplanned workload is most essential and who ought to work on it. This avoids scope creep and distractions that often occur when unplanned work doesn’t get managed systematically.
The assessment course of boils all the way down to 4 major actions:
This leaves us with the problem of accountability (or the dearth thereof). To plan for unplanned work we have to outline roles and priorities, in order that duties are literally taken care of in a well timed method. We remedy the accountability downside with Triage tasks. That is an explicitly outlined particular person liable for coping with incoming requests within the Triage queue.
If we follow our 911 analogy, this particular person is successfully the primary responder.
You possibly can manually choose one or a number of group members who will get notified (or auto-assigned) for every new difficulty, or use an present PagerDuty schedule to automate the method.
Lastly, we will use service-level agreements (SLAs) to speak urgency and set steering on completion deadlines. SLAs are utilized mechanically relying on pre-defined variables. For instance, you possibly can set particular timelines for pressing bugs or for requests out of your most essential prospects.
SLAs make it easier to to ascertain and preserve inside requirements for the way rapidly unplanned work will get resolved. However additionally they give your group a framework for prioritization, in order that an important points and requests get resolved first.
Outro
Wrapping up, let’s summarize an important factors and offer you some helpful hyperlinks so you possibly can arrange your personal plan for unplanned work.
- Unplanned work are bugs, alerts, and different requests and emergencies that seem instantly. You possibly can’t escape this sort of work, however you possibly can handle it.
- Unplanned work has two broad phases: An consumption section and an execution section.
- The consumption section is about turning reported bugs and requests into express, actionable points. To make this course of as frictionless as potential, you allow consumption proper the place unplanned work first seems — with integrations to your most essential instruments. Linear has integrations with buyer help platforms (Intercom, Zendesk, Front, Plain), engineering instruments (Sentry, Incident.io), and many other products.
- Slack is a very essential consumption channel the place loads of unplanned work surfaces. We built a tool called Linear Asks that enables your group to rapidly flip these requests into actionable points.
- Within the execution section, we first centralize all incoming unplanned work in a shared group inbox referred to as Triage, the place every difficulty will get reviewed, prioritized, and assigned. You can learn more about Triage and how to enable it for your team here.
- To get probably the most out of Triage, we suggest having a clearly defined person who reviews incoming work. For bigger groups, enabling SLAs may help with the prioritization of time-sensitive requests.