Start to Finish relationships in Project Management

Seriously, I don’t think I’ve seen such a simple topic cause so much confusion in the project management community.  I guess one of the principle causes for confusion lies in the lack of practical examples, so let’s clear it up by providing some practical examples of start to finish relationships.

What is a start to finish relationship?

In simple terms one task cannot finish until another has started.

That’s it.

When trying to find practical examples most people get confused by trying to make finish to start examples fit.  I can see why as in both relationships the tasks are linked at the same points, the logic is reversed- follow the arrows.

Finish to Start

In the Finish to Start (default relationship)  Task 2 can only start once Task 1 has finished.  It’s actually hard to think of two related tasks that don’t share this logic.  By following the arrows you can see the flow from the Finish of task 1 to the Start of task 2.

 Start to Finish

The difference with Start to Finish relationships is that a task cannot Finish until another Starts.  Simple in theory, but not too many related tasks follow this logic.  By following the arrows you can see the flow from the Start of task 2 to the Finish of task 1.

You may also notice something quite critical.  In the Finish to Start image (the blue one) the arrow has been stretched to say that there is no urgency to continue the work. In this example once task 1 is finished we will start task 2 the next day (as per our working hours).

However in the Start to Finish image (the orange one) you can see a small arrow and task 2 is being stretched over the non-working time (shaded area).  This is a critical difference of this relationship.  Task 1 cannot finish until Task 2 starts, think of a baton relay it basically needs a real time exchange- task 2 starts and only at that moment task 1 can finish.

 Real world Start to Finish relationships

The Start to Finish relationship can be found where continuity is required between two tasks.  Basically at some point one of the tasks needs to be occurring.

I have had students aim to justify answers based upon the linking of tasks such as delivering a cake and billing a customer, having a party and cleaning up.  In most cases they have been confused by trying to force the use of a start to finish relationship on tasks that should actually follow a Finish to Start relationships.

Shift Work

Probably the most obvious example.

Shift A starts at 9am and is scheduled to finish at midday.

Shift B is scheduled to start at midday and finish at 5pm.

Now I’m not trying to talk you out of the Start to Finish relationship but consider the other options.

Can both shift happens at the same time, this may be an obvious no.  But lets assume they were 2 teams doing data entry. Providing the workload could be split and sufficient resources were available we could get away without linking the tasks at all, allowing the 6 hours of work to be completed in a 3 hour duration.  The plus side is that since there is no link, if one shift was late it would not affect the other.

The next best link using the Shift Work example is the Finish to Start.  Using this relationship Shift 1 would start at 9am once they have finished their work this would allow shift 1 to start.  If shift 1 finished early, this would allow shift 2 to also start early.  If shift 2 came in late it would not affect shift 1- they could go home once complete.

Now for our Start to Finish relationship.  I would only recommend using this if there is an actual requirement to always have a shift on site.  For example lifeguards at a public pool, if the pool would have to be shutdown if there were no lifeguards on shift than the Start to Finish relationship is suitable.  Further examples include a control room operator on a gas facility, without this role the plant could not continue to operate.  A nurse in the emergency department of a hospital and so on.  The main way this relationship affects the schedule is that if shift 2 is late, shift 1 must continue to work until shift 2 finally start work.

So whilst the start to finish relationship could be used for shift work, it should only be used if there is a requirement for continuity; i.e to always have a shift working.

Continuous Systems

This should further help drive the point home.  If there is a requirement to have continual operation it makes sense to use the Start to Finish relationship.  In the following cases Task 1 normally represents the operation of the existing system and Task 2 is the new system.

Here’s some examples:

Video Recording– if there is a requirement to have continual video recording of an area- such as a within a city council.  The recording performed by the existing camera could only finish once the new camera had started to record.

Data Backup– if we are replacing the physical backup for a computer system which could not be shutdown, such as a server.  The old system would need to continue to operate until the new system had started.

Changing power supply– Getting bigger now and adding a little complexity. We have laid new cabling infrastructure in a region and as part of the project we are going to link a hospital to the new power system.  Due to the life sustaining equipment utilised in the hospital there must be a constant power supply.

So a Start to Finish relationship makes sense, we can’t finish using the old power supply until we have started on the new one.  However what happens when we start the new system and it fails?  In high risk situations there would be a redundancy, in these cases normally a backup generator, which kicks upon the detection of power loss.  But let’s assume we didn’t have that luxury.

We can build in a scheduling redundancy by adding lag.  So we still use a Start to Finish relationship, but the logic is task 1 can only finish after a day of task 2 starting.  This allows us to ensure that the power supply is stable.

Lag is a useful accompaniment of the Start to Finish link for the reduction of operation risks as seen in these scenarios as well as often being an occupational requirement; i.e handover for shift workers etc.

 Activity interaction dependence

Ok so the last category of examples I’ll write about are when two activities are linked, but there is only a small acceptable window between the finish of the first task to the commencement of the second.  Cases where quality maybe affected if the time between the finish of one task and its successor is too long.

Concrete

We are getting concrete delivered to site by a contractor as soon as it arrives the next task is to pour the concrete to form the foundations.  Using Australian Standards we have about 90 minutes from mix (normally done prior to the truck leaving) to pour.

The two tasks we will be focusing on are

• deliver concrete,
• pour concrete.

If I asked you to link these two tasks 99% of people would give me something like this picture, which is fine for now.

Now, I’ve also added a couple of extra tasks- just to give it some real world feel.

The logic is pretty self-explanatory, in order to pour the concrete we also need to have

• Completed the form work and
• Obtained the approval for the delivery truck to enter the site

Have a go at drawing up a schedule for these activities.

1. Completed the form work (3 hrs)
2. Obtained the approval for the delivery truck to enter the site (10 mins)
3. Deliver Concrete (1 hr)
4. Pour Concrete (10 mins)

Check out the Start to Finish Example article to see the answer

 

Now I hope this makes sense, best of luck with your schedules and remember don’t use the Start to Finish relationship if you don’t need to.

Happy to answer any questions as well as provide examples that may be of relevance to a particular industry.