Gantt charts, work plans and funding research

I spend four hours yesterday making a chart. Really.

It was a Gantt chart. For those of you not familiar with a Gantt chart this is what they look like:
Gantt chart

According to Wikipedia – source of all modern knowledge – these charts were used in the first world war, designed by Henry Gantt (1910-1915). I am not sure why it took Mr. Gantt four years to design said chart, it is after all a simple creature the chart. Equally, while I am sure that WWI was an organizational nightmare, I am not sure we need to hold a WWI chart up as the pinacle of management organization. First of all there wasn’t much movement in the first world war; just a lot of death and destruction and pointless bloody battles. So the organization must have been largely a static one. Not to mention nobody actually *won* the first world war – in the sense most of us think of winning a war. The German Army collapased. So why did I spend four hours making a (slightly modernized) version of a WWI chart?

Funding agency requirements and a really great administrator who knows what I need to do to meet said funding requirements. (and no I am not being sarcastic – he is great). My first version was rejected – too vague. My second version still, too vague. My third version (I hope) passed the mark.

I think it is good that as academics we have to justify our time, set goals and show that we have an idea of how to organize a research group. This is essential. If someone is going to give you a sum of money to pursue research, that they want to see that you can organize said research in some kind of reasonable way is more than fair. But frankly, using a WWI organizational chart to do this, isn’t the best idea.

Scientific research is hard to put in a chart. First of all, some experiments take more time (or less) to accomplish than you assume they will – and you can’t really predict this. It is like replacing the wallpaper in a Victorian age house. You look at said wall and think, I can strip that in about 3 days, tops. Emboldened with your steamer you set about your work, then you realize that all of the plaster board is falling off. They you realize that the wall was actually patched together by someone in the 1960s with newspaper and glue. 3 day job turns into 3 week job.

This is (often) what happens with scientific research. Looking on paper at a project you say – I can do these experiments in about 3 weeks. Ok, this is not so bad, you probably can get the experiments done in three weeks but you waht you can’t say is I will understand all of the data in exactly 6 months time or I will write a research paper on this subject starting October 5, 2014 at 2 pm. But this is precisely what Henry Gantt is asking me to do. Science doesn’t work like this. It is not a static moving around of munitions but really a dynamic process. You do experiment X then often what you think next is WTF? So you have to go do about 20 more experiments to figure out what your favourite special experiment X (which you just KNEW was going to work) is actually telling you. Or maybe you think you know what the data is telling you but it is not so definitive so you have to do 20 more experiments to really know.

Now this is not to say that scientists have NO idea what is going on. Most of us know what initial experiments to run and techniques that will work in a particular instance – but what you don’t know is what the results are before you run the actual experiment; and often you dont’ get a definitive answer from said experiment.

So while I agree that explaining your work plan to get funded is a great idea, showing some kind of management savvy is essential; can we please get leave the Gantt chart back at the turn of last century where it belongs.

About Sylvia McLain

Girl, Interrupting aka Dr. Sylvia McLain is a bio-physicist in the Department of Biochemistry, University of Oxford (UK), but she blogs in a personal capacity. She is also a proto-science writer, armchair philosopher, amateur plumber and wanna-be film-critic. You can follow her on Twitter @girlinterruptin
This entry was posted in research funding and tagged , . Bookmark the permalink.

23 Responses to Gantt charts, work plans and funding research

  1. Anon says:

    I’m an active researcher & put Gantt charts in all my grants, even when not required. I love them. It is a much more efficient way to convey the timing info than in words. You can build in time for changes of plan & write up etc. and even I you never stick to the timing for real, it is useful to formally think about time planning.

    • fair enough- I don’t mind justifying myself but I don’t think they are always the most useful mechanism – especially for short projects …

      • Seconded – I think Gantt charts and flowcharts should be compulsory in all grants, as they can really clarify the thinking behind the research plan to the reviewers (and often to the applicant(s), too!) You don’t usually have to stick to them any more than you would have to stick to the original research plan if an unexpected result in your first year of funding leads you off in an exciting new direction (depending on the funding agency and the rules of the particular funding competition, of course)

        • yes I am getting that – I have had my ass kicked about charts – I would say that is true for big grants – in fact in reality I do do it – It was just this one was v small – for 1 PDRA who did 3 tasks – I did have a chart, it just wasn’t good enough – maybe I just don’t understand them yet ;(

  2. Hi Sylvia
    I’m another nerdy person who rather likes Gantt charts. Of course you can’t always predict accurately, but they can be good for highlighting when starting task B depends on completion of task A – in which case, you can get an idea of whether the whole project is going to be borked if task A takes longer than expected.
    When I review grants, I’m surprised to find instances where people really haven’t planned it out realistically. E.g. they want to recruit 250 children and then test them 3 times over 2 years, but they don’t take into account the fact that it’s likely to take them at least 6 months to recruit the initial cohort. If they’d plotted it out on a Gantt chart, then it would be clear that they were going to run out of time, but they didn’t and fail to see that the project is doomed. The alternative that you also get is that they might say, well, it is going to take 6 months to recruit, and THEN we will start testing, but in fact the testing could start before recruiting was complete, so the overall time would be less. And sometimes you can see that they haven’t thought through who will be doing what – either they’ll have people sitting around doing nothing waiting for something to happen (such as ethics approval) or they may have a sudden spike in work that they won’t be able to cope with. It can get especially hairy in longitudinal studies where people need to be seen at particular time-points, and a good old Gantt can be a useful visual aid.
    D

    • I can see the point – especially for big grants – where they are useful. But for small ones (like the one in this case for this blog) I have 3 tasks 1 PDRA for a pretty small amount of money. I think they aren’t always the only way, but yet they are always required. You can also eloquently describe what you are trying to do in bullet points and still organize accordingly.

  3. Ugh. I’ve done many of these things, as we have some funding agencies and other schemes that absolutely require them. I can see the point for helping to lay out and clarify for the reviewer what exactly is planned, and how long it will take. What I can’t see the point of is expecting periodic progress reports tied to the milestones and objectives in the chart – because things change in science, all the time. In my field, technology is moving so fast as to be continuously disruptive, and so as to render milestones like completion of a certain number of experiments almost meaningless. New discoveries by competitors or collaborators can change plans. A new postdoc showing up with a fellowship can radically alter the lab’s productivity. And so on, and on.

    The other problem I see with your chart is that you seem to be using Microsoft Project, which is The Work Of The Devil(TM). ;)

  4. no way did I use Microsoft anything – just stole a picture from the web …
    I am a Unix girl – I abhorr windows with a passion ask my group ;)

  5. Cloud says:

    I manage a group doing scientific informatics in industry. All of my projects have project plans (which is all a Gantt chart is- a graphical representation of a project plan), but the more researchy projects and the projects in which we are delivering enterprise software have very different sorts of project plans. I don’t personally find the graphical representation of a Gantt chart to be the best way for me to track my projects, but I still make them because they are how senior management likes to see overviews of the plan. Also, I know some very good project managers who prefer Gantt charts to the format I use (which is called a work-breakdown structure). I think that choice is a matter of preference, not substance.

    I have heard you argument about how it is impossible to make a detailed plan/Gantt chart for research many times. I completely disagree. Without knowing how far out and in how much detail your funding agency required a plan, I can’t say whether or not its particular requirements are silly. But I can say that I’ve made plans for many, many research projects, and they have been useful. Research projects have decision points, and experiments that you will do to get to the decision points. When I make a plan for a research project, the milestones are the decision points, and the experiments are some of the tasks. Importantly, the work required to prepare for an experiment- such as producing and purifying protein for assays, writing code to analyze data, or buying/building the equipment needed to run an experiment- are also tasks, and I note the dependencies. This helps me identify risks to my timeline and makes sure that forgotten dependencies don’t introduce unnecessary delays. It also makes sure that I know who is going to do the work for each step, and helps me keep an eye on how much work each of my team members has, so that no one gets overloaded. I find it useful to plan 2-3 decision points ahead in detail, with the other decision points left high level until I know whether we are going to get to them.

    There are risks and uncertainties in the other fields that successfully use project plans. When I make a plan for a software development project, I have just as much uncertainty as when I make a plan for a research project. The risks that lead to that uncertainty are different, but they are still there. Research science requires a different sort of project planning, but project planning can still be very effective, and done well, can help you get more research done in less time.

    • I don’t think it is impossible just that often they aren’t useful – what you said in the first paragraph I think sums it up – ‘I personally don’t find a Gantt chart the best wasy for me to track my projects’
      I also agree project planning is very effective and you need to do it, but I don’t think a chart or diagram always works – and often really doesn’t work. What I am opposed to is being required to do it in an absolute way. Not that you can’t do it for science or even you shouldn’t do it for research. I think you should plan – I just don’t think the Gantt chart works, and don’t like having to HAVE to use a specific chart. Why can’t you use what you want to plan as long as you show you can plan – does it matter if you use a diagram or bullet points or just words? I think it shouldn’t.

      • also see Richard’s comment above – I think that sums it up

      • Cloud says:

        Why don’t you think that the way the people making the funding decisions like to view the plans has any importance? They have to look at lots and lots of these. Why should they have to learn your specific quirky way to present your project plan? I may have a different format I use for tracking my project internally, but when I present to upper management, I put the plan in the format they like- not just because they are the bosses but because they need to make decisions across projects, and need a uniform way of assessing project progress.

        But honestly, I think continuing this discussion would just waste both of our time, so I’ll stop. I’m not going to convince you to modify your view.

        • Easy now – i am not totally disagreeing with you …. in fact I am agreeing with part of what you are saying ….
          I am not saying what they want has no importance either, not at all – just that it doesn’t always make sense – and that trying to uniformly do everything like this may not be the best way.

          if a proposal is being assessed mostly on the science, where they have to read the science – shouldn’t the way you decide how to orgnaize be dependent on the science ?

          Just having a friendly discussion – for my part

  6. Steve Jones says:

    WWI was an industrial law – I rather suspect the use of gantt charts was in the background logistics and industrial productivity rather than planning battles. Like their cousins, PERT and Critical Path Analysis they aren’t too good at dealing with rapidly changing circumstances in “real time”. I rather think that the German amy didn’t rely on these techniques for bltizkrieg (although they may have paid for it – apparently German WWII offences often ran out of steam when their logistics failed – rather different to the post D-day alied offences, where logistics were everything.

    • Laurence Cox says:

      PERT was developed in the 1950s by the US Navy, originally for the Polaris programme. If you have a deadline (let us say a conference paper submission date) then it allows you to identify the critical path (the path from the project start to the project finish that is the longest and hence tells you the latest time you have to start the project) and also where elsewhere in the project you have free float and can move activities around to make best use of your resouces. It is a really useful tool for all project planning and not to be sneered at. Gantt charts are one of its standard output formats, and one that is most used for illustrating the progress of the project. As long as users appreciate that it is not something just for impressing grant-giving bodies and should be regularly updated throughout a project as part of the project management process, then it is a very useful tool for identifying if you are on-track and, if not, what you can do about it.

  7. Great comment @Laurence – this makes sense to me. and also @deevybee – I can see that example and how it works well – I really am starting from all of your comments to see the use.

    In this particular case, which I was writing about – I am only reporting on something very simple – eg 1 PDRA 2 tasks – once they complete task one they move on to task 2 – hence my irritation for filling out a chart for this. I don’t think it is always needed. I don’t always think it is appropriate and I definitely think they shouldn’t be actually required when not needed.

    But I do think keep track of things and updating things (and timing things) is important. It is something I sort of do anyway – eg keep track and things in a big book (minus the formalized chart). I think maybe the discrepancy and what I find useless is when it is ‘required’ when it is not needed. Maybe I just don’t know how to use it.?

  8. John Finn says:

    Nice post!
    I think it was Eisenhower said “Plans are useless, but planning is essential”.
    I used to feel the same about Gannt as you, but have changed my mind. They can convey planning very well- the real trick is for funding agencies to take Eisenhower’s view of plans and planning! We’ve never been tied to the Gannt chart in a proposal, thankfully.
    As a grant reviewer, Gannt charts are also very useful and important to do well ( effective and well thought out , rather than fancy).
    Regards, John

  9. rpg says:

    It’s good to have a plan. Otherwise what else are you going to abandon when your ideas meet reality?

  10. Jonathan says:

    Out of curiosity, do you have a favourite package for producing Gantt charts under Unix? I’ve looked around for something to use in latex documents, but struggled to find anything that was easy to use and looked pretty.

    I was very skeptical about Gantt charts at first, but they do force you to work out an ordering for tasks – a little like writing computer code.