Mixed cluster and stack charts

On a few occasions, I had to use a combination of a cluster and a stack chart. This chart is not available as a standard option in PowerPoint. Here is how to make it:

• Create a regular stacked column chart
• Set the gap width to 0
• Blank out the data where you want the gap between the years to be
• Manually add labels for each of the years

You can create one yourself using the above ingredients, or you can download the one I made in the SlideMagic template store:

Consistency in financials

Financial projections of new business ideas are totally made up / not accurate, so being of by a few million here and there would not matter? We can make quick changes in our financials in the presentation slides, and then "forget" about updating our financials spreadsheet with the new information. 

While the absolute numbers of your financial model might be totally pulled out of the hat, it is the thought process of how you got to them that is still valuable for investors. How does your business model work? What would I have to believe in order for your year-5-dream-scenario to come true?

And that model should be consistent across all your documents: presentations, spreadsheets, budgets, everything:

• Discrepancies make you look sloppy (a little preview of things to come when you need to work together with an investor on a Board)
• A consistent model of totally made up numbers makes sure that everything is, well, consistent. If you just slashed sales & marketing cost by 50% but maintain the same amount of sales people, something goes wrong.
• Inconsistencies make it harder to understand your story for an outsider. If sales are $50m on one page and $49m on another people get confused. You established "$50m" as a mental shortcut for year-5-sales-in-the-most-optimistic-scenario, and all of a sudden you create a new shortcut.

So, even if nobody can predict the future accurately. there is still value to create a consistent financial model the same way as you would for a business in which you know every single detail (next year's budget of an established company for example).

What can you do to incorporate the fact that numbers are highly uncertain?

• Round things up to whole numbers (no $49.569m sales in year 5)
• Minimise the number of assumptions you put in the model and make cells that are guesses highly visible (I usually mark them bright yellow).  Rather than "guessing" the number of customers for each year, and the number sales people for each year (10 assumptions over 5 years), you could assume a % growth of customers, and a fixed ratio of sales people to customers (2 assumptions).

Slightly more complex models might actually be simpler to understand.

The unit measure

There are endless ways to show financial ratios and benchmarks. Each industry has their own specific ones.

Beer brewers think of the manufacturing cost of beer in terms of cost per hecto liter (not liter, not kilo), but when they think about distribution cost, they think per 6 pack case (not liter, not kilo, not bottle). The cost of sending a letter is usually in weight, the cost of sending a package is usually expressed in terms of volume. Retailers think in terms of sales per square meter. Technology startups think often in terms of per user per month.

Find out what the norm is in your industry and adjust your financials.

Tedious number updates

I am not a believer in copy-pasting Excel numbers into PowerPoint. Excel is for analysis, PowerPoint for presenting, and presenting numbers is different from analyzing them. The charts I use to present are often so simple (only one data series, rounded digits) that I tap them in by hand.

There is an exception though. If you are the junior analyst on the team, and you are the last step in the "supply chain", i.e., constantly updating charts with new versions of numbers that roll out of the spreadsheet, it might be time to change working practice. Personally, I remember many McKinsey projects I did as an analyst where I designed a model in week 1 of a project, and ended up updating numbers on slides for weeks after that.

In these cases, add a worksheet to your Excel file, and pull the numbers from the analysis. Apply all rounding, i.e., if the chart or table requires millions, divide by 1e6. For tables, put all numbers in the right order, invert rows, columns if required.

You can go further and actually create objects in Excel that can almost literally be copied into PowerPoint. Create data charts, fix fonts, colors, etc. exactly as in the PowerPoint file. Format Excel cells exactly as the table in your PowerPoint document.

The cost of this effort is time: it will take some effort to get it right, but it will pay off in subsequent number update rounds, and it can prevent you from making mistakes in last minute model changes.

Be careful not to introduce new mistakes though. This type of work can best be done when things are relatively quiet. Take an existing, stable version of the deck, and start recreating the charts in Excel, constantly double checking whether you can replicate the numbers you put in by hand in the charts. Once you are confident that everything works for an old/trusted version of the document, you can use it for future versions. Still, some spot sanity checks are always a good idea.

---

Image via WikiPedia

Chart chooser

There are endless types of data charts out there: lines, bars, scatters, pies. Which one to pick depends on the type of data you have, and what message you want to convey. Here is a new handy tool that can make life easier for you: chart chooser.

A box of handy cards that guide you step-by-step through a selection process. Excel templates are an optional add-on that could be useful, some of these diagrams are tricky to recreate.

Investment bank memos

Investment bankers usually prepare the data rooms for due diligence in M&A transactions. Shelves full of information that is usually summarized in an information memorandum. I never found these summaries very useful. It is impossible to create a financial picture in hour head from just reading text with data. "Last year's sales were x, growth for the 3rd quarter was y, the company is doing business in 3 product segments." The only way to understand this is to extract the information and put it in tables or graphs to see what is going on.

And that is most of the times exactly what the buying side will do, hire a consultant/analyst to go through the material and start creating these charts, then someone more senior will go through the charts and select which ones are important, and which ones not.

You can accelerate the sales process by anticipating all of this, and do some of the homework of the buy side. In the process, you can spoon feed them the right insights you want them to have.

Similar but unrelated, this is also why wordy descriptions of company results in newspaper articles do not work for me. The journalist took the data tables and graphs and translated them into words, the exact opposite direction of where I want her to go.

---

Image by Andreas Poike on Flickr

Visualizing contrasts

This article on Vox tries to find statistics on voting fraud in the US. I will stay out of politics on this blog, but the 2 graphs it uses show the power of a good visualization.

A column chart shows the relative difference, but fails to communicate the large overall absolute value of the right column

The image does a better job to depict the statistics. It requires some math and Photoshop skills though to produce

The column chart is the correct representation of the 2 values, but it fails to communicate the huge amount of ballots we are talking about in the right column. The image does a better job, but it will be hard to construct for a layman designer.

---

Image from WikiPedia

One science is harder to explain than another

Most of my client work involves a presentation with a hard/difficult business or scientific issue to explain. All science is complex, and it requires someone to study for years to understand it. But in some scientific disciplines I can get away with things.

Take medicine for example. Here is what you can do to make things understandable to the layman:

• Micro focus on one very specific condition/disease, and omit all other 35,000 medical issues a student usually has to go through
• Simplify, eliminate complex / long / Latin names
• Abandon commonly used diagrams, and instead make your own, completely non-standard drawings that are solely aimed at explaining a phenomenon.
• Shrink all the statistical proof into a footnote. It is important for peer-reviewed research, it is not needed to understand the basic mechanism of a drug

This can work in other disciplines as well: economics for example. But it does not work everywhere. Mathematics for examples requires a broad understanding of concepts that are all stacked on top of each other, depend on each other. And there is no easy to simplify formulas. 

---

Image from WikiPedia

Words and number consistency

It is impossible to make a correct 5 year business forecast in an investor pitch. But your financial projection is not really a forecast, a prediction of the future, it is a picture made out of numbers. "If our company will be successful, this is what it could look like".

Mistake number one is to make the projection ultra precise with 5 digits after the comma. It is just a guess, so a year 5 revenue number of $99,234,318 is not more credible than ~$100m.

But oversimplifying is not right either. The fact that you are for sure going to be wrong does not mean that you simply take 1% of a big market number to get to the year 5 scenario.

The trick is to make the words/visuals in your presentation consistent with your financial model. You are going to sell to millions of individual customers, drive the model that way. You rely on 5 big telco operators, put it in. SAAS company with recurring revenues? Model it. One-off perpetual licenses, use it as the basis of your model.

Teaching investors how the business works is more important than getting the point estimate right.

Grasping data charts

Here is what a first time viewer needs to go through when looking at a data chart:

• What's the unit on the y-axis
• What's on the x-axis
• What do the legend colours stand for
• What is "good" or "bad", are things supposed to go up or down?
• What is actually happening
• Is the y-axis broken or does it start on zero?
• Why is this unexpected, interesting, counter-intuitive, amazing, terrifying?

If you did the analysis then you live and breathe the data in the chart, the only thing you watch out for is how the line/column/bar looks different from the previous version of the chart you saw. For a completely new audience it is a different story.

There are 2 types of data charts:

1. Charts to ponder and refine research and analysis
2. Charts to communicate the result of the pondering and refining

If you make a presentation of result, you deal with scenario 2. Make a diagram completely from scratch. Start with "what is actually happening" and only put that piece of data on the chart.

For example if a complex stacked line diagram shows that the % of returning visitors gets smaller over the past 5 years, you can replace all that monthly visitor data with 3 data points: % of return visitors in the last 3 years.

---

Image by Phil Roeder on Flickr

Storyfying financials

Most presentations have some sort of written summary on the first page. You often almost skip it when presenting live, but it can be important for an important that reads your slide deck at the computer rather than attending a live presentation.

The one thing that I find really hard to summarise are financials. Prose that goes like "sales were x, up y% compared to last quarter but z% versus the same quarter last year at a slightly higher operating margin of v%" makes me totally lose the picture.

Financial journalists fill pages of the Wall Street Journal with this type of texts. And they are upset that it is now possible for an automated bot to do this job: insert a financial table and out comes a perfectly written paragraph. (Robo journalism)

What do I do as a reader? I reconstruct the table back in my head. 

So what to do in a presentation? I recommend a super simple written summary on page 1 (sales were up, but profits down), and then a very crisp financial summary table on the next page. Round up numbers, and use colour coding to show what went up or down.

---

Image from WikiPedia

US maps with statistics

The recently joined web site Data USA (datausa.io) is a great source of maps that can serve as backgrounds for presentation slides.

Overdoing the animations

This animation (GIF alert) shows the distribution of music sales over time. Wait a few seconds and you see the pie chart changing for multiple years. This data can be represented much better by a series of stacked column charts. The animation takes too long, and the audience does not have the overview of all the years. 

I copied this image from a Tweet that did not include the reference to the source.

There are other problems as well with this chart, gradients, standard Microsoft Office colours, drop shadows, small data labels, and ambiguous labelling ("Internet", "mobile", "video", etc.). 

Combining tables and data charts

Lining up a data chart and a table in PowerPoint or Keynote is very tricky. And that is a shame, because it is one of the most useful compositions to present data. Just tables, and you cannot really see the trends. Just data charts, and it all becomes cluttered.

I took the data from an earlier blog post and quickly turned it into a combined table/data chart. You can clone the slides I create in presentation app SlideMagic into your own SlideMagic account by clicking this link.

Table or bar chart?

Back in the good old days at McKinsey, we tend to put any range of data in a column or bar chart, even if that meant a chart with just one column, or one chart.

A nice clean bar/column chart works beautifully to visualise a data range. Especially if there are big differences between the values of the chart. If the differences between the numbers is not that big (I do not like broken axes), or you have lots and lots of data to present (the data labels just become too small), I resort back to a simple table. 

Stick all the data in, round things up nicely, and use accent/shading background colour to make your messages pop.

I must admit, I start using tables more and more. (And that is why they are so easy to make in SlideMagic)

---

Art: Louis-Philippe opening the Galerie des Batailles, 10 June 1837 (painted by François-Joseph Heim)

Confusing ranges

A large of the data presented in business presentations is uncertain: future forecasts, market (share) estimates, competitive comparisons. Many people express that uncertainty by presenting numbers as ranges: cost per bottle as between $1.50 and $2.00.

While working with ranges might be required in a spreadsheet calculation, I don't like to use them in presentations:

• It clutters the slides with double the amount of data labels, and/or double the amount of text in tables
• It becomes hard to work with totals, add things up. Five of these 1.5 to 2 becomes 7.5 to 10. Adding negative numbers into the equation will give totals that nobody can work out from the top of their head.

So what to do? I would stick the mid point of the ranges as a point estimate in the presentation, and makes sure it is a nice, rounded number (7.49544 looks to certain). Much simpler and much clearer.

---

Art: a painting depicting women inspecting silk, early 12th century, ink and colour on silk, by Emperor Huizong of Song.

Sales presentations can be technical presentations, and vice versa

Often, clients tell me that they have to present to an audience of engineers, so the presentation should not be a "marketing" presentation.

The bad interpretation of this is: we are engineers presenting to engineers, so we can get away with text-heavy slides and diagrams full of numbers. We engineers understand each other. As soon as we add pictures or try to make the presentation more visual in other ways, we lose credibility.

What is really going on is this: if you are selling a high tech product there is no way you can avoid well, talking about the technology. But the high tech world is full of presentations written by people who do not really understand the technology, for an audience who does not really understand the technology. These presentations lack substance but are rich in marketing buzzwords. Engineers will recognise them in a second, and don't want you bring "one of those".

We need to eliminate two misconceptions:

• Technical presentation content cannot sell
• A senior (and/or) sales presentation audience does not understand technology

Here are some of the things I do to create technical sales presentations:

• Insist on explaining how it works in human language without "black boxes" or "secret sauces". The Einstein quote about a 6 year old who can understand everything if told in the right way applies here.
• Very focussed data visualisation. Technology advantages are often beautifully simple: things are faster, cheaper, smaller. Rather than writing a bullet point "we are 33% smaller", put in that very complicated data chart that shows the full richness of your research, but add 2 big bold lines that are 33% apart. 

Keeping things up to date

Some presentations contain a ton of data that needs to be updated all the time. Quarterly results, LP updates of VC/PE funds, the latest sales data. Updating the numbers is time consuming and errors can easily sneak in (especially problematic with presentations to investors).

I do not recommend cutting and pasting Excel data into PowerPoint. You need serious PowerPoint skills to format the data correctly, and most spreadsheets are not build to present data, they are build to analyse it. Hence my approach of cutting the link between the spreadsheet and the presentation software, and creating a data chart from scratch, 100% focused on the audience, not the analyst.

How to deal with the updating?

I would suggest to create a special spreadsheet alongside your presentation. A new worksheet pulls the required numbers from the big "data dump" worksheet, rounds them up correctly. Place the numbers exactly as they should show up in your presentation slide. Now it is easy to update your presentation data. Add check sums to see whether percentages add up to 100%, and breakdowns go back to the total sales figure.

When the new quarter arrives, you over-write the data dump worksheet, and fix any broken/misplaced links.

---

Art: painting by Ivan Aivazovsky

How to make a sankey diagram in PowerPoint

Sankey diagrams can be useful to show flows.

They are tricky to make in PowerPoint. The width of the arrow needs to correspond with the value of the stream. The curves of 90 degree arrows in PowerPoint are hard to control. If there is no escaping (maybe you can create a waterfall diagram instead), I create Sankey diagrams using boxes and triangles, see the example below.

---

UPDATE: I have added a PowerPoint template for a Sankey diagram in the SlideMagic template store:

---

Art: Minard's classic diagram of Napoleon's invasion of Russia, using the feature now named after Sankey.

Cleaning up survey results

Quantitative market research companies come back with pages and pages and pages of data in PowerPoint. It is tempting to cut and paste to "add the survey results in the presentation". Not so fast, let's clean things up first.

1. Find out what the research actually says. This is the most important step. There is meaning in 2 levels. Level 1 is to discover what trends the data displays (not always obvious from a randomly generated bar chart or a dense table of numbers). Then level 2: is this an important insight? A segmentation or cut is not always meaningful, sometimes data points do not actually vary that much. Sometimes conclusions are obvious / not surprising, not interesting.
2. Ask for access to the original research data in a spreadsheet to find patterns, trends that have not been put in the final report of the research company
3. Clean up the graphs you want to use (or even better, create them from scratch with just the data you need):
   1. Move non-essential statistical lingo/jargon/details in small print to the foot note: n values, standard deviations, etc. etc. are not worth the screen real estate in a stand up presentation. The person who wants to read them, will find them in the foot note. (Exception: certain scientific disciplines where one statistical value is all that matters, in that case make a very prominent chart with just that value.)
   2. Cut non-essential filler words from the data series labels, think as if you were writing headlines for a newspaper article or blog post. Simplify questions that were asked to respondents, put the original questions in the footnote or in an appendix chart.
   3. Re-sort data series to match the point you want to make. 
   4. Take out tick marks, make bars/columns fatter, take out Excel labels and titles, and insert your own in PowerPoint/Keynote, round numbers, fix chart colours to match your corporate colour scheme
   5. Where necessary, add circles or arrows to highlight the point you want to make. Adjust the colours to emphasise your points (apply accent colours correctly)

---

Art: Edward Burne-Jones, The Mirror of Venus, 1875