A guide to poster making for scientific conferences

Jun 10, 2019 13 min read

Source Document

How to make a good poster
Recipe
Content
Design

This is a tutorial on how to make a scientific poster from scratch.

It aims for first-time poster makers, but also for people like me that quickly lose themselves in choosing the right font for hours.

It includes a few self-declared best practices that helped me in the past. Those are at best ideas, and barely opinions. It is by no means a collection of strict rules and feedback is welcome!

I got lots of tips from Flavia Hodel during stimulating discussions around poster presentation. Flavia also gave valueable feedback after reading through a draft.

How to make a good poster

An excellent video by Mike Morrison recently made the rounds online. If you have not done so already, watch it before continuing.

Mike Morrison makes the point that most of the posters presented at conferences are not easy to grasp.

There are two principal reasons for this:

The content is not adapted for a poster presentation. For example reusing whole paragraphs from a manuscript.
Poor design choices. For example small fonts, problematic colour choices, low-resolution figures.

Separating content from design is important^[If you know Markdown or LaTex, you know what I mean.] because it will allow you to focus only on one or the other. And since “design” is a profession, a better description in what we are doing is “polishing” or “improving readability”.

I will structure my tips into these two parts: content and design, along with a recipe that I like to follow.

But let’s talk first about what makes a poster presentation special.

Oral versus poster presentation

If you get an oral presentation (congrats!), then you know that your voice will carry most of the talk. While your slides should be carefully designed, you can still rescue imperfect slides with a good talk.

Poster presentations are somewhat special because they are at an intersection of a talk and an art exhibition. Meaning, you will once during the conference have the opportunity to guide other researchers through your poster. This is similar to a lightning talk, just more engaging.

The remainder of the conference, your poster will hang there, and maybe some people will still try to understand your topic. No matter how good you are at talking, now the poster has to be self explainable.

In fact, your poster should read like going through stand-alone slides (slides with more explanations to replace the speakers’ voice).

Back to the video

While overall the video of Mike Morrison is excellent, I struggle with three things.

First, it talks about emphasising this very clear conclusion. In real life, this is often not possible. For example, because a research project has just started. So instead of having this one research question, I’d go for the main message - what do you want that people remember? That can be the large sample size of your study or that you are developing a new method.

Second, the poster template still crams the details of the study into a small space, which is not ideal for readability.

Third, it talks about a horizontal^[Update: As Peter Higgins pointed out, horizontal is common in the US, vertical is common in Europe.] poster, and this is often not possible at conferences, due to space constraints^[However, a straight forward solution is, to turn the design by 90 degrees.].

Recipe

Here is a recipe that I like to follow when I make a poster.

Get ready

Start at least 7 days before the poster needs to go into print^[Make sure you know how much time the printing service needs to print your poster.].
Think of the main message: what is the one thing that you want the reader to remember? Write that message into a text file.
Take a pen and sketch a draft of your poster on an A4 paper. What do you want to present and where does this content go on the poster? What kind of figures and tables to you want the reader to see? No need to write proper sentences, instead use keywords and skeletons of figures and tables. You can also do this step at a later stage, basically whatever suits you best for your creative process. For example, you might want to do first steps 4-7 before sketching a draft.

Get your abstract and extract title, author and affiliation and save it into a text file too.
Create a folder img somewhere on your computer.
Copy all text, tables and figures you possibly need into the img folder. Slides from previous talks can be a handy source.
Get the logos of your university and affiliations and store them into the img folder too. If you have the choice, go for PNGs, as they have transparent backgrounds.
Choose a tool that is easy to operate and does the things you want. My choice is keynote because I cannot bother with Adobe Illustrator. More details below.
Update: Read the conference instructions, particularly on poster size and orientation (as recommended by Peter Higgins).
Resize document to the size recommended by the conference or A0 (841 mm x 1189 mm or 2384 pt x 3370 pt). This way, even though the poster might be proportional to A4, you can use the real size fonts.
Split the document into different boxes (or any compartmentation according to your sketch). For example four horizontal lines and one vertical lines will turn 10 boxes. Use rulers, guides or lines to help you with that. Each box will later answer a different question.

Now you are ready to add content.

Add content

Before adding content, choose a sans-serif font. Helvetica or Arial will do a good job. These fonts were carefully designed to be easy to read and look good on print. Later on, you can still decide to invest time into checkout google fonts, but remember - it needs to be sans-serif. The smallest text size should be 24 pt.
Fill the top rectangles.
Place the title on top. This is the first thing the reader should see.
Add authors and affiliation right below.
Add the logos next to the authors and affiliations.
Next, add your main message below.
Add contact and lab webpage at the very bottom of the poster.
No need to adjust and fiddle around with boxes at this stage. Just fill the content.
Now we are left with 7 boxes. How you distribute the content into these boxes depends on your topic. I work in a field where data and methodology is central, so I always dedicate one box to data and one box to methods.
The first box will take care of the introduction and should answer the question Why is your research needed?
The second box should present the methods: What method did you use to answer your research question?
The third box takes care of data presentation: What kind of data did you use and how does it fit together with the method?
The fourth and fifth box are reserved for results: What are the results if you combine data and methods?
The six box is dedicated to a discussion or summary: What can you conclude from the results? What are the limitations? And where are you heading next?
The seventh box is for references and abbreviations.
Update: Add a small portrait of yours to the upper corners or next to your contact email address. This way people can spot and talk to you outside the poster session (recommended by Federico Marini).
Update: Add some empty post-it notes and a pencil next to your poster and ask readers for feedback (recommended by Federico Marini).

Polishing & Design

Now that you have all your content, you can start to improve the readability.

In order to quickly grasp a poster, it should have as less sentences as possible.

Consider replacing some text with icons or figures.
Make use of bullet points.
Align all text and shapes (again, rulers might help here).
Make sure all section titles are equally sized.
Are your figures (and the rest of your poster) colour-blind friendly?
You can start adding colours (if they serve a purpose).
Think about whether your figures need captions or not.

Finalising

See how it looks like on a printed A4 version (update: or for less troublesome reading print it on A3, as mentioned by Michael MacAskill).
Ask your colleagues for their opinion by showing them the printed version.
You might want to let it sit for a day or two. Then look at it with fresh and rested eyes.

Trick-trick: Have a detailed and a lean version

After a few days break you look at it again and you realise that it is still to busy with text.

Here is the trick - keep two versions: the detailed one you already made (that can be accessed through a QR code), and the one that you will bring to the poster session with lot less details.

Make a copy of your poster draft, name it poster_detailed.pdf. Put this pdf somewhere online (e.g. dropbox, google drive) and get a link to it.
Generate a QR-code of that link and download the png to your img/ folder.
Continue working on your original file.
Count your words. Your aim is now to cut the number of words in half - roughly. If you don’t know where to remove content, present your poster to a colleague from the same faculty (not the same lab). You will quickly realise over what details you jump. Remove those details.
If you can simplify the graphs - do that too.
Count the words again - has it been halfed?
Add the QR-code to the bottom of the poster.

Printing

Print it again in A4 for yourself and check:

if all text and shapes are aligned and equally distributed,
if the font is the same throughout the poster,
if all authors or affiliations are present,
for typos,
if figures are clearly readable (and in high resolution),
where the QR-code leads you,
whether your contact email address is correct.

That’s it - send it to the printing service!

Content

Before adding any content, think about your audience. These will be scientists, and they have probably heard about research you are working on, but don’t know the details and motivation behind your work. In fact, there will only be a handful of people that know exactly what your domain, topic and method does. All others are still interested though. So lets aim for these people as your audience.

After looking at your poster, the reader should have an idea what the title means. Your poster title might be super clear to you and your lab colleagues, but anyone else will have a hard time understanding right away what you are doing.

Therefore, one goal is, to clarify the words in your title.

After separating your file with rulers as described above, you will have 8 boxes, of which you can use 6 for:

Introduction
Method
Data
General results
Specific results
Discussion & Summary

I recommend to give your boxes titles so the reader can jump from one to the other.

What content should go in your boxes?

I believe that a poster cannot carry a condensed version of a full manuscript. I think it is ok to focus on only one or two threads of your work. Your story telling needs to be coherent - that is all the reader wants.

Introduction

I like to follow the paper writing guidelines by Jennifer Widom as a rough guide:

What is the problem?
Why is it interesting and important?
Why is it hard?
Why hasn’t it been solved before?
What are the key components of my approach and results?

Methods

Unless your very focus is a new statistical methods, keep this section simple and clear.
Make sure that it is clear why the method is able to answer your research question.
You can use equations, but simplify them too, or annotate them with arrows and explain all fancy letters used.

Data

Mention the origin(s) of your data.
What was the sample size?
No need to describe all the variables you used. To save space, you can also group them into larger categories.
Describe distinct features of your data that could influence the interpretation: e.g. if all your samples were male.
Mention how you processed the data. Again, no need to be specific. If your area uses standard QC methods, mention that the data was processed according to standard QC techniques.

Results

Have two layers of detail. For example, go from general to specific; have a graph that summarises your results, then zoom into your results in a next graph.
Use figures to illustrate your results and avoid tables.
Simplify the figures and use the power of your poster making tool to annotate the figures.
Make sure that the results are follow the notation of data and methodology. E.g. if you present results for a trend in age, make sure you have mentioned age before.

Discussion

This is the place to:

summarise your work
conclude
talk about limitations
announce future plans

Contact

Add your email address.
Add lab webpage.
Add the QR code that leads to your (detailed) poster.

References

Add important references, for example software that you used.
Make sure you cite original content.

Design

Or improving readability.

Tooling

A good tool should satisfy the following criteria:

Decent in zooming in and out.
Variety of graphical options.
Operable with ease.
Something you can use over years.

If you have practice with Adobe Illustrator or its’ open-source counterpart Inkscape - go for it!

Otherwise, google slides, Powerpoint, keynote - although limiting - do a the job too.

Fonts

No matter what sophisticated typeface you had in mind, choose one of these sans-serif ones for a start: Helvetica or Arial.
If you really have the time to play with other sans-serif fonts, check out google fonts.
Making fonts is an actual job - remember to pay for special fonts.
Most tools have a replace font option, so you can replace a font through out your poster.

Icons

Some icon services are for free, or at least partially: e.g. flat icon.

Figures

Principles for data visualisations could easily occupy another blog post. There is much to say, but let’s keep it short here.

If you use R and ggplot2 for making your figures, use one of the black and white themes, e.g. theme_set(theme_linedraw()).

Make sure that the axis text size is also 24pt once it is on the poster.

Colours

Colours are either used as decoration or to encode a variable. Do not use colour for your background^[Black or gray background and white text colour might work though.].

Always use colour-blind friendly colours.

You can test the colour-blind friendlyness of your poster and figures with https://colororacle.org, a colour-blindness simulator on your computer, or in a browser: https://www.color-blindness.com/coblis-color-blindness-simulator/.

For Figures:

If you use more colours than available in a discrete colour palette - drop that encoding. No one will be able to grasp.
Use colour-blind friendly palettes from the colorblindr R-package.

Resolution of figures

Your principal figures should always be stored as PDF format. This way, you can easily turn them into a high-resolution PNG.

To turn a PDF into a high-resolution raster image, use convert from imagemagick.

If your file is called figure-1.pdf, then write:

convert -quality 100 -background white -alpha background -compress lzw -units pixelsperinch -resize "789x2625" -density 300 figure-1.pdf figure-1b.png

If you have lots of pdf starting with figure-*, then write:

for file in figure-*.pdf; do \
echo $file;\
convert -quality 100 -background white -alpha background -compress lzw -flatten -units pixelsperinch -resize "789x2625" -density 300 $file `echo $file|cut -f1 -d'.'`.png;\
done

Full disclosure

Here are some of my posters from different stages of my career. Making posters is a process of improvement.

2018

2017

science communication conference