Last year, we wrote a detailed, subjective ranking of sellside research note graphic design. The impact was swift: after a bottom-half ranking, Credit Suisse collapsed within six weeks.
Whispers have also reached FT Alphaville that our review (which, let’s be honest, was mainly just grousing about Arial) caused a stir — prompting discussions in some research departments over how to improve the presentation of their output. Which is good, because several of the notes were poorly-designed to the point of actually reducing usability.
But we’ve also been plagued by doubts.
As most readers know, Chamath Palihapitiya is FTAV’s north star. Rather than being afraid of our own shadows, was it time for us to enter the arena? Was it time for us to try to design the perfect sellside note?
In the name of content, yes it was.
As most readers will also know, Alphaville sometimes already does sellside notes (which some might say look a little like Deutsche Bank’s), so let’s keep the same loose premise. We’re Alphaville LLC, and we produce research on basically anything.
Let’s get into it. Graphic design, we suppose, is just a series of decisions.
Decision -1: What is perfection?
We stipulated that our note ranking was subjective because:
— Duh
Obviously there’s no right-or-wrong way to do graphic design. Everyone has different opinions (and, more importantly, responses) on colours, typefaces, layout, the presentation of data, information quantities etc.
It is worth looking over what’s already out there. Here are some of the example notes we looked at, to establish the loose vibe of the format (aka do market research):
There is also, of course, also academic literature, which can furnish us with certain quantifiable understandings about how people in aggregate respond to certain visual elements, and can help us understand things like best practices for accessibility.
So we’ll try some of that, and ignore it when it suits us. 🕺
Decision 0: What do our clients want?
Ach. Well the last thing we want to do is actually ask them. So let’s make some assumptions. We reckon users of sellside research are after notes with:
— lots of clearly-presented information . . .
— . . . but that can be quickly skimmed if needed
— layouts that minimally interrupt the flow of their reading
— charts!
Decision 1: How big is a piece of paper?
It blows our hivemind that the United States doesn’t use the ISO 216 paper size standard (A4, A5 and the gang).
Like, we consider ourselves worldly people and are aware of America’s little idiosyncrasies like mass incarceration, the widespread availability of assault weapons and not being able to transfer money via your banking app, but come on — look how absolutely great it is to be European:
Instead, Americans prostrate themselves to bizarrely-named paper types of seemingly random size: Letter, Legal, Tabloid (Ledger) and all other types of sordid nonsense. We’re not even going to include a picture because this is a family-friendly finance blog.
Point is, we won’t stand for it: FTAV may be intercontinental, but we have standards. Our note shall be A4. Behold:
Decision 2: What colour is our note?
There’s a temptation to make our notes look a lot like the Financial Times, the airport freesheet which likes to associate itself with Alphaville.
The FT describes its signature paper/background colour (#FFF1E5 online) as pink, but it’s also kinda orange. We say probably, because perception of colour is a situational and there are a lot of different situations.
Here’s something we can say for sure: the A4 paper that Alphaville LLC’s clients are buying is neither pink, nor orange. It is white.
Asking our treasured clients to buy pinkyorange paper, or to print a pinkyorange background on our notes, would be inefficient and wasteful. So, white background:
Rapid progress.
Decision 3: OK — other than white, what colour is our note?
The 60/30/10 rule is sometimes thrown around in graphic and interior design: it suggests that 60 per cent of a colour palette should be the primary colour, 30 per cent a secondary colour and 10 per cent an “accent colour”. Why not?
With this in mind, our rough thinking on usage is something like:
— 60: Main heading, body text, numbers
— 30: Some subheadings, labels and links
— 10: Drip
Our 60 is obviously going to be black — or, rather, very dark grey (#333333), because black on white creates an unnecessarily extreme level of contrast that can cause legibility issues for some users:
The other colours are tougher to choose. There are also a lot of options, so we decided, for branding reasons, to raid MainFT.
There’s a darker version of the FT’s signature pinkyorange, #FCD0B1, that is used for the logo. If it makes no sense for ol’ pinkyorange to be our background colour, could we use this variant as our primary accent?
Absolutely not. #FCD0B1 has a bad contrast ratio with white, according to both Web Content Accessibility Guidelines and the evidence of our eyes. It would look rubbish:
Where do we go instead? There are plenty of places that suggest humans link certain colours strongly with certain topics, but the reasoning is usually extremely janky, relying on a combination of pseudo-evolutionary vibes and Western-centred assumptions.
So, uh, let’s be guided entirely be accessibility and distinction. The sellside notes we looked at last year primarily (no pun intended on any level) deployed reds, blues and greens as secondary colours.
Do we want to stand out from the crowd? Yes, of course. But how? Here are our options, from the FT’s standard palette:
Let’s try, uh, bright claret (#ff1493). It’s pinker than many of the reds of our competitors, while also nodding a bit to ‘FT pink’ and has the best contrast with white out of the non-gloomy options available:
Could Alphaville LLC’s future be hot pink? We’re willing to try.
For accent colour — which we reckon could work for boxes and that — we decided to keep the two conventional FT pinkyoranges. Branding, innit?
Decision 4: The fonts
It’s far enough down the article that we can be honest with each other: one of the first things the author did upon joining the FT was download its font files and email them to himself so that, were he sacked, he’d still have copies to play around with (ed note: this is the saddest thing I’ve ever read).
And, once again, we’re going to get high on our own supply — echoing how most sellside research design loosely follows broader corporate brand guidelines. The FT has nice typography, and the studio behind most of it — New Zealand’s KlimType Foundry — has a fantastic stable of fonts (yes, Arial Bold is a font and Arial is the typeface, but semantics are boring).
Across print and online, the FT uses five main typefaces, with distinct use cases:
Sell-side notes tend to come as PDFs, and the printed document in the form we’ve found ourselves naturally drifting towards. But is research normally read on a screen, or in print?
We’d err towards the former, which would suggest that Georgia is our best body font option if we want to use a serif font . . . which, we do, because Georgia works better than anything else here. It’s also a standard system font on Windows and Macs, which is a big advantage because it means it should also render properly:
So what about headings and other accoutrements? Alphaville LLC’s instinct was to go for Financier Display, but it’s maybe too FT. It makes things look like a newspaper, rather than research (which tends to have more modernist design leanings).
Metric, however, does seems to work:
Finally, we’ll roll out Univers Light Condensed if we needed it for areas where space is at a premium. If it’s good enough for Goldman and Deutsche, it’s good enough for us (not a general rule).
Decision 5: The grid
The perfect book doesn’t ex —
OK, OK, we’re all Müller-Brockmann stans here, let’s get swissy with it.
If you have no idea what we’re talking about, grids are basically an underlying layout system commonly used in web and print design. Here are some examples:
Now, one of our favourite note formats is Standard Chartered’s. It’s not the flashiest, but it does the basics really, really well: the text isn’t too wide, and charts are presented in line with when they are mentioned in text (something Deutsche also does quite well by having a sidecar column). Our praise drew this interesting comment from user WillSG on last year’s piece:
Grids matter. Most of the notes we looked at last year used five-column or three-column grids, although Deutsche deploys a very fancy seven-column model.
Which to choose? There’s always an argument in favour of simplicity, BUT given this is just a framework to put text into, that shouldn’t usually be a problem for the end user of the template itself.
So let’s go with . . . seven, which offers maximum flexibility. Maybe we’ll regret this.
We’ll aim to have our body copy span five columns, and whatever we want over the other two. It will let us break up the top of the note (where there tends to be more random stuff like logos):
Decision 6: The logo
Alphaville has never had a conventional logo, so it’s not clear what Alphaville LLC should use. Let’s just do a diamond made from an A on top of a V, which represents both our name and our diamond hands 💎. This is almost certainly already someone’s logo, but (whisper it) we’re not a real LLC:
Decision 7: The layout
There some key elements most sellside notes seem to share, roughly in order of appearance:
— logo
— company/division name
— a key information panel with the sector/date etc
— title (with a price target nearby)
— a contact information panel
— a company information panel
— a summary of the note
— the main text
— subheadings
— charts
— tables
— disclaimers
Within those, there are myriad possibilities and associated questions. Let’s focus on the bread-and-butter bit first: the note itself:
Do we want our clear spaces to be left or right of the text? StanChart goes left with summaries, Deutsche right with charts:
StanChart, in our view, works better because summaries are always small, whereas you might want to have several charts or a wide chart, necessitating extra columns. Let’s create a test layout using rough StanChart style — imagine the grey square below is a chart:
This . . . feels OK‽ We played around with going it on a 5-col grid as well, but it felt like there wasn’t enough space for text in a 2col/3text layout.
We’re worried Georgia is looking a bit fusty, but part of the issue might be the Latin lorem ipsum text we are using. Metric would probably look more handsome, but then we’d have potential accessibility issues if our PDFs encounter glitches. The alternative would be to go back to step 4 and expand our options beyond the FT’s font stable, but the number of decisions that would involve scares us.
How about the top bit? Banks seem to vary between a ribbon splash of their secondary colour across the top (eg Nomura, Numis and Rabobank), or a header section that is separated by a small line or a big block with the header (eg BofA, Credit Suisse, HSBC).
We’re aware that Alphaville LLC’s hot pink could be quite striking, so let’s try to get just a bit of it into the header:
It’s OK. Where’s the rizz though? Put it in wider context, and the whole thing looks a bit anaemic:
One, uh, rizzy style that we like is TS Lombard, although their notes seem to have calmed down a bit since last year:
Could we find a middle ground between the two for our headers? Maybe a zany underline? We tried a few different things, including experimenting with Financier Display — despite having written it off as too newspapery earlier:
There are a lot of ways this could go. First things though — we’ve aligned the title with the text, ie it starts from the third column of the seven. This pushes us towards a contact/company information panel on the left, á la Peel Hunt.
There’s an argument that this is practically useful (a typical right-handed reader can hold the note in their left hand without covering the core content), though on-screen that’s irrelevant.
The title type itself though . . . it’s tough. There’s a very real danger of ending up looking too bloodless. Maybe some top-heaviness would help? We tried hot pink, but it was way too much. The FT pink, however, seemed a bit better. We also put in an actual company name, rather than a placeholder:
White space is lovely and important, but we also had information to convey. Let’s get a summary, analyst details and the rating information in there (the latter could be dropped for economics research or replaced with a disclaimer saying all economics is vibes).
Joyfully, Financier Display has an entire set of associated dingbats, we could raid for a nice, hefty bullet point:
For the other parts, we just dropped in some filler content and used AI to whip up a generic-sounding UK equities enthusiast:
Trimmings added, we were nearly there:
Decision 8: Dataviz
A picture speaks a thousand words, while a chart typically speaks between perhaps five and five thousand. This wasn’t the toughest decision — we ripped off the usual FT style, but dropped in our hot pink. If we needed further colours we’d raid the paper’s palette as needed.
Aah:
Thankfully, it looks a lot calmer in CMYK. After some further tweaks, our note example was ready.
At this point we decided to also make an economic variant, to show the style’s ✨versatility✨. Losing the pricing information freed up space to the left of the title, which we’ve used for a cheeky lil flag. There’s a risk this looks a bit tacky but it does make the note easier to identify at a glance.
Some other useful things for the rest of the column . . . [side]footnotes, useful images or a calendar? We went for the latter two.
Et voilà:
Some inconsistencies have emerged, particularly around the spacing under the titles. After staring at our screen for too long, we decided we had been staring at our screen for too long.
And, sure, there are other things to do (dense back-page data tables and disclaimers, mainly), but the fun bits were done. We’d merged together some of the key things we felt were important:
— distinct branding ✔️
— in-line charts ✔️
— paragraph summaries ✔️
— clarity . . . hopefully ✔️
But creation is only the first step of the process. It was time to gather feedback. And who better to do so from than you, our dear readers?
If you have views on the design, comment below or email your thoughts/ideas/abuse to Louis, and we’ll return soon for some reiteration. And if that doesn’t get you excited, we don’t know what would.
