How far have you progressed in your actuarial career? And what skills do you need right now?
Whether you're a graduate student or a fellow, there's one skill you're guaranteed to need: writing well. Whether it's an exam paper or a group actuarial report, you'll be judged on how clearly and concisely it's written.
Anyone can learn to write well
Considering the importance of writing well, why are so few actuaries taught how to do it? Is it because 'anyone' can learn on the job and if you don't, it's your fault? Because studying the art of writing is for humanities graduates, not actuaries? Because writing is some mysterious gift bestowed on the few? Because either you're a numbers person or a words person?
All these 'becauses' are myths. Yes, a little artistry helps, but much of business writing is like carpentry: master the tools and you can build solid documents with seamless joinery and plenty of polish.
Know your outcomes and readers
As with carpentry, you first need to clarify your aims. So just as it's not 'planks on a wall' but 'somewhere to put your books', it's not a 'report for management' but 'persuading management to do X not Y'. Before you write a word, ask yourself:
- Who's my audience?
- What do they want or need?
- Why should they care?
The top reason documents fail is that they don't address these questions. For example, a policy full of passive sentences ('reports are written, risks are monitored') has failed to achieve its fundamental purpose of explaining who must do what.
The question 'Why should they care?' is crucial, because people read only when they have to or when they want to. It's naïve to assume that 'have to' is enough to get people to read. Think of those terms and conditions that are so boring we click 'Accept' without reading them: for all we know, we've unwittingly agreed to sell Google or Amazon our kidneys.
How do you discover what your readers need and want? It sounds nugatory, but talk to them. What they care about and know (and, critically, don't know) is likely to differ wildly from what you assume. Smart questions to ask readers include:
- What do you need to know? Never assume. When a colleague of mine was asked to write a report for 'Exco' he naturally assumed his readers were the exec committee. He realised too late that it was the board, so his document failed.
- When do you need to know it? People don't want 'a document', they want information - so rather than focusing on delivering an artefact, focus on delivering the right knowledge to the right people at the right time. The medium is not the message.
- What's relevant to you? For example, both the board and actuarial team may need to understand their firm's catastrophe exposure - but the board is unlikely to care about (or even understand) the implications of modelling using a Gaussian versus Student-t copula.
- How familiar are you with the subject? This question affects both what you say and how you say it. Continuing the previous example, all the board may need to know is that the Student-t copula is preferable to the Gaussian copula because it better meets regulations about accounting for the lack of diversification under extreme scenarios but it increases the SCR by £10m.
- What do you need to achieve? For example, a car manual for mechanics (how does it work?) is completely different from one for drivers (how do I use it?). Understanding objectives also helps determine tone of voice: 'I want to assemble this flat-pack furniture (so I'm cool with being told what to do - but don't be too curt!)'; 'I want to know how to make a complaint (so don't patronise me because I'm annoyed enough already).'
- What's your background? English as a first or second language (so go easy on the jargon and slang)? Colleague or client (so steer the right side of casual/formal)? Senior or junior (be deferential upwards, but kind and supportive downwards)? And so on.
Gather your materials
Now you understand your audience, what next? Gather your materials, which can be either primary (what you know or can ask) or secondary (what's written already).
It's a waste of time regurgitating secondary sources - what I call 'rewriting the wheel' - so ask yourself:
- Is it documented already? Any time spent searching your intranet and asking around is more than recouped if you save yourself the bother of writing yet another internal document nobody needs. And if information exists in the public domain, a URL is a lot less trouble than writing it yourself - actuarial glossaries and regulation requirements being classic examples.
- Can it be generated? For example, if you need a list of documents with their metadata (owners, authors, review dates, etc), there's no point creating it manually in a spreadsheet when SharePoint libraries do it for you - including giving you an 'Export to Excel' option if you need a point-in-time status report.
Select or design a template
The best templates are easy to use and legible: clear typeface, optimum line length (40-70 characters, roughly eight to 12 words), ragged right margin (more legible than justified), and headers and footers fed by fields such as filename and 'last saved' (easier than having to hand-type the doc name and version number each time).
Pick a filename
A good filename tells readers two important things: what the document's about, and its relationship to other documents. So follow a consistent filename taxonomy, such as Client_topic_doc-type.
Build the structure
It's wise to create a skeleton of headings with a clear narrative spine before fleshing out each section. Give each 'bone' in your skeleton a meaningful label (not 'Introduction' but 'Introduction to <whatever>') and apply a structure strong enough to support your text: for example, familiar to unfamiliar, generic to specific, chronological, and so on.
Write the first draft
When writing your first draft, a great trick is to structure each section according to what journalists call the 'inverted triangle' - heading, followed by the 'lead' (a one or two-sentence summary of what the section's about) and only after that the details. Information's far easier to grasp if it's contextualised first - and heading/lead/body aids skim-reading.
Don't get too hung up on finessing your text when you're bashing out your first draft. One secret of great writing is 'write fast, edit slowly' - so you unleash your right-brain creativity with your nagging inner critic muzzled, then let your left-brain critic rip your draft to shreds.
Edit, edit, edit
After your first draft it's all about the editing, which includes:
- Grouping information into meaningful chunks using intelligent paragraphing, tables and (in lowest position of usefulness) bulleted lists.
- Creating smooth transitions between each section, paragraph and sentence, and even within sentences. Ways to do this include using transition words showing how one idea relates to the next, such as whether it's a result (therefore, so), an example (for instance), a contrast (however), the next in a sequence (first, second), and so on. Another way is using what's called 'referential continuity', which is lining up sentences like dominoes, with the end of one matching the start of the next to form a logical sequence.
- Removing ambiguities and unearthing buried assumptions. One way to do this is stress-testing your sentences by asking whether they address the 'five Ws and one H' (who, what, why, where, when and how).
- Trimming verbosity, and fine-tuning your authorial voice.
There's much more to it than this, of course, but I've run out of words so you'll have to wait for the next article. Which reminds me that another good writing technique is to finish on a cliffhanger
Julian Maynard-Smithis a senior technical author - international operations Axis Capital www.betterbusinesswriting.info
