Results at a point in time, e.g. a survey (ref 53)

First, a warning: casual readers might think today's email repeats last month's. No. Last month's looked at how to structure and format tables, and showed a graph over time. Today's is, counter-intuitively, about adding stuff to tables, and its graph looks at data at a point in time, not over time.

Jon’s Number Nightmares: we’ve five restaurants (Green, Red, Brown, White, Blue), and people have given scores for their food, décor, staff, in-house music. High score is good. See Figure 1. White’s figures are all along on one row. But we’ve a problem. When we see White’s Staff score is 76%, we ask ourselves: “How does that compare to others?”. So we reorder the Staff column. In our head. A bit of a pain. And to see how White's 53% Music score compares, we reorder that column too.

Try Figure 2. Each column is in descending order – and here’s the counterintuitive bit. Yes, there’s more typography – White appears four times. And yes, it now takes a little longer to find White’s numbers – they’re no longer in a row, so we must scan the table to find them.

But yes, the table is better. OK, it takes a second or two to find White’s numbers, but we then instantly grasp its position within the distribution. Is it near the top? Or the bottom? We get insight. We no longer need to mentally reorder a column of numbers. The benefit of grasping numbers a lot more quickly far outweighs the benefit of finding numbers a little more quickly.

Of course, this assumes that up is always good – if we've a column of, say, Complaints, invert it to aid at-a-glance comparability.

More on graphs (or - perhaps - 'moron graphs'): maybe we should put the numbers in a graph. Try Figure 3. We saw something like it last month, and it's awful.

What about Figure 4? OK, at least columns are in descending order. But we don’t engage with them. We glance at them (“Gosh, columns get smaller…”). I once saw a benchmarking report with 60 page-filling column charts, spread over 60 pages. Terrible.

So try Figure 5. It puts the restaurant names/logos in four silos. It doesn’t look like a number-cruncher’s graph. It starts at ‘50’ to help differences stand out – see bottom left (and I’ve highlighted ‘50’ – I don’t want people to think I’ve sneakily truncated the axis in order to deceive).

Look at Food. Tightly clustered. The rest are more spaced out. Also, see how Green scores well for three out of four, but its Staff score is low. We spot patterns. We could, of course, spot them in the graphs above. But we don’t. We glance at them.

Logo chart or four columns of numbers? Which is best? It depends. The tables surrender detail first and patterns after. The logo chart surrenders patterns first and detail after (to find exact scores, we must refer to axes, make estimates, etc). Also, with the logo chart, there can be problems: logos sometimes overlap. Not ideal, but not a disaster either.

Back to the dining table: let’s improve it further. Averages give extra insight - see the top row of Figure 6. Which helps highlight that columns are in random order, and last month, I said random order is bad.

Study Figure 7. Columns are in order. There's an Overall column too. Compared to the table we started with, Figure 7 has lots more in it. More labels, columns, rows, numbers. But it's better, we engage with it much more easily. (Of course, these changes can also be made to the logo chart: (A) add an average label to each silo; (B) put silos in order; (C) have a Total silo.)

Note also the following four points:

(1) Maybe embolden a restaurant's name - it helps if we, say, work in White's diner.

(2) Follow these tips intelligently, not slavishly. For instance, in sports league tables, only one column is in order: the points to date. That’s because it has primacy.

(3) What if we’ve 20 restaurants? No worries - I've done similar tables for 23 companies. Logo charts struggle with that many though. (What if we've 120 restaurants...? That’s for another day.)

(4) Remember to eschew frippery (shading, funky fonts): Figure 8 is how Excel encourages us to show numbers. Small wonder people think they aren’t good with tables – for years, we’ve struggled with drivel such as this.

What about an infographic? Maybe we could show a picture of a kitchen table (table… geddit?). And show Music scores on a stave and Food scores on a stove. Ha!! Word-play. Visual gags. Love it!! Then again, maybe not.

Next is the token joke bit, but first here's a postscript on the 60-graph benchmarking report I mentioned earlier. It was a yearly report, but one year, its authors were busy, so I was asked to do it. I got rid of every graph – and as I sent bosses my graph-free report, I held my breath. You see, the report went to the biggest bosses in a big company. Some liked detail, some liked overview, and none were slow to complain if they disliked something. The outcome? The report never returned to its original authors. I remained its author from then on.

The jokey bit: this month's email had lots of Figures about food. So here's one last one... With thanks to my brother for sending it to me many years ago.

On that note, let's finish.

PS for a macro that creates the logo chart, click here, enter details, and it’s download 020-1.

Clarity and Impact Ltd | +44 20 8840 4507 | |

To see more of my previous email updates, click here.

Want to sign up for my monthly emails? Click here.

Want my GDPR policy? Click here. It's a bit irreverent, plus has two jokes.

Clarity and Impact Ltd