Non-fluffy startup strategy

How I create and communicate a non-fluffy company strategy that my team can understand, and can use as a decision-making tool

This is the first (proper) post in my newsletter about my deliberate approach to learning to be a CEO, and I hope a good one. The second post, now published, is about writing investor, board and team updates. You’ll find that, and more, here on the archive.

The varying success of Lingumi over time has roughly mapped to the quality of our strategy.

During the couple of years we spent working without a clear strategy, we didn’t drive growth or find product-market fit. In the subsequent year and a half, we have had a clear, non-fluffy strategy that we’ve followed rigorously, and that has led us through the initial product-market fit and into the early growth stage, where we are as I write this.

This is by no means pure, single-variable causation, but a non-fluffy strategy is table stakes: you won’t win just with a good strategy, but you will lose without one (unless ‘win’ means Google acqui-hiring you 😛).

My co-founder and I now walk every new joiner through this journey, detailing those years of scraping around for product-market fit in a rather aimless fashion, and the changes we made as we ‘became more strategic’ (buzzword bingo, sorry), and it’s the highlight of the onboarding process for us, serving as a cautionary tale, and as a cornerstone for future work.

Since then I’ve told a few peers about the way I create and communicate our strategy, and it seems to strike a chord. Instead of endlessly repeating myself, I thought I’d write it down. I’m not wedded to it - there may be far better expressions of strategy, but this one works for us right now.

I have a tendency to be a ‘trench fighter’, ticking off tasks and being in the weeds of work. To me, ‘writing a strategy’ in 2016/2017 would have felt like lazing around or slacking off (‘there’s way too much to do!’). I now believe that getting less ‘sh*t done’, and spending much more time preparing a non-fluffy strategy for the team to get behind and work from, is somewhere in the top 1 highest leverage things I can contribute as CEO.

An early warning: if you’ve read Richard Rumelt’s Good Strategy, Bad Strategy (or this snappy summary of it - really, leave this piece and read it! Thank me later!), a lot of this will be familiar. My friend and former colleague George changed my life, and Lingumi’s path, when he lent me his copy, and most of the work we do at Lingumi on strategy is directly taken from Rumelt. However, application, consistency, and rigour are my main tools here, rather than the core concepts, and I am strict about how I apply them. Focus and rigour in a single direction are the key takeaways, not ‘make a fancy slide’. I hope you enjoy reading about it. Oh, and if you’re a superfan, you can share this post too <3


How do you format your strategy?

A friend and fellow CEO, Kate, recently asked me ‘what should a strategy slide look like for a board deck? Should it include all our KPIs or only some?’

I don’t know what it should look like, but my strategy slide doesn’t include any KPIs at all. Our Lingumi strategy slides have two very simple lines: a problem statement, and a guiding policy. The team gets them, the board gets them (…and they don’t get a window-dressed version of the problem to make things look pretty). They’re very simple and communicate their point in as few words as possible.

A strategy is a way of overcoming a problem. Though a startup might have lots of problems, if I spend time and energy just thinking (which I’ve learned that I should do much more of - it is real work!), I can usually distill Lingumi’s most fundamental problem(s) down into 1-2 sentences. There are lots of fires to fight: here I try to outline the biggest fire, or better, the root-cause of the big fire. This is the problem statement.

The strategy at Lingumi is an articulation of this problem, including a simple set of “guardrails” indicating the direction to travel in order to overcome it. This latter part is the guiding policy. Again I use 1-2 sentences, and it’s not a task, it’s a signpost.

That’s it: 2-4 lines, broken into two parts. That’s our strategy, on one slide.

Our best strategies have been very ‘un-fluffy’. They have helped us say ‘no’ to things, more than helping us decide what to actually do - your team is probably best placed to figure that out, or experiment in pursuit of what that is. My worst strategy slides were too fluffy, or the direction of travel wasn’t explicit enough, because I was too weak-willed or lacked the self-discipline to push aside all the wonderful ideas floating around, and wanted to leave chinks in the armour to allow them to slip through. I’m constantly trying to be narrower, more focused, and more explicit with these strategy slides, and struggle against my own bad habits to want to ‘build all the ideas’.

Enough theorising! Show me the goods (and the bads)!

Fine. Let me give you some real examples. The first is how to do it terribly. The second is how to do it better.

This was my first attempt to apply the ‘Problem Statement / Guiding Policy’ combination, c.August 2018. Strap in for some seriously fluffy strategy:

Ok so let’s analyse it...

Problem statement: actually not too bad. Good job, former me. This is clearly a problem: we were trapped in flat linear growth that was becoming too expensive (as everyone else who relied in Facebook in 2018 was...but at least we acknowledged it!).

Product Guiding Policy: this one seems pretty bifurcated, you’re saying we work on both child and parent experience with a team of 8? This isn’t focused, this doesn’t give an unambiguous direction to the company.

Market Guiding Policy: easy tiger…two guiding policies? Call the fluff police👮‍♀️🚨 And this one’s even worse! It’s saying ‘be good at direct marketing’ (what does that mean? Did we have those skills?) and be good at getting academics on board, and sign partners...oh yeah, all with a team of 8.

What a clusterf*ck of a strategy, no wonder we had a bad Q4 that year! I feel embarrassed sharing it, but this is good group therapy, thanks readers. It was aimless, pointed in many directions, and didn’t help reduce decision-making anxieties or increase singular focus. Luckily, our strategy was much stronger in Q2/Q3 2019 and we began to drive affordable, meaningful growth after a total product rebuild, focused on building a better product for our core customers, shutting off non-core markets, and growing the elements of the product that had inherent network effects. We scraped through the year, and began to grow fast in mid-2019.

Now for a better example. First, the context:

At the end of Q4 2019, we were gearing up to publish the third level of our first course on the Lingumi platform. We’d released a successful re-vamp of our first two levels (which we totally focused on, in lieu of building new features - good strategy!), and customers, pleasingly, were finishing those and crying out for more. Revenue and retention depend upon us continuing to build out the course, so the logical step would be to build level 3, ASAP, right?

Well, yes, right to an extent. But after much agonising thought and discussion, we made this our strategy for the quarter:

We did have a major problem down the funnel: customers crying out for more levels! And boy are those super users noisy. But we had a bigger, silent problem higher up: new customers not converting to super users, because of issues around stability, user experience, and generally low ‘trust building’ in the onboarding (where we ask for mic and camera permissions). We’d been hearing these as a quiet drumbeat from our CS team, and seeing it in our data (or rather, seeing the outcomes of it), but were always looking at the new, shiny things customers wanted, and not focusing enough on the ‘boring bits’ around stability and trust.

Both of these potential strategies would have been ‘the right thing for the customer’ - it was just different types of customer, at different stages in the funnel, with different needs. We couldn’t do everything, so we had to look ahead, asking ourselves ‘what will break as we scale?’, ‘what are the brutal truths?’, and ‘how patient are the customers waiting for L3?’. The decision was agonisingly difficult to make, but it was a good strategy because we made a firm decision and focused.

The Guiding Policy set the whole team on fixing these problems across a customer’s lifetime, and across segments (so some team members focused on new users, and others on older users deeper in the platform). We didn’t touch level 3 all quarter.

Critique time: how could I have improved this?

I think the Guiding Policy is a bit fluffy. What does ‘across customer segments’ mean? I should have written ‘for both new leads and older customers’ and defined two categories that we knew we had to work on. The policy above still leaves this a bit ambiguous. Secondly, what does ‘develop’ actually mean? I could have totally cut this, and just focused on ‘activate’ and ‘retain’. I am often too wordy, and it’s visible here.

I am trying not to give ‘advice’, but here are anecdotes and thoughts for making the most of this approach to strategy if you do want to use it:

Clear Guiding Policies reduce decision anxiety!

During Q1 2020, our guiding policy reduced decision-making anxiety, helping us make the right decisions. Our VP Engineering approached me during the quarter, telling me that a stability fix needed another two weeks of work if we wanted to ship it, but that’d delay his work on Level 3. What should he do?

The answer? Check the Guiding Policy. They spent two more weeks fixing it. If the guiding policy isn’t clear enough to answer major ‘A or B’ questions where A and B are radically different pieces of work, the policy is probably too fluffy.

How do I test if the strategy is good?

Our strategy’s first job is to align the team very tightly. Even if it’s a wrong strategy for the business, it can be a good strategy if the team know it and can move faster and in a united way from it. If the strategy is succinct and non-fluffy, you can ask anyone in your team ‘what’s our strategy this quarter?’ and they ought to be able to give you, perhaps not verbatim, a run down of the problem statement and guiding policy.

Test 1: I ask a colleague ‘do you remember what our strategy is?’ with a coy smile, explaining that if not, it’s my fault, and they should be very honest about having forgotten / not absorbed it.

Our strategy’s second job is to reduce decision-making anxiety, but without removing space for debate or innovative work. In other words, it should tell you which direction to point in (and make it easy to say no to pots of gold under other rainbows), but should tell you team what to actually do.

Test 2: Ask a team member, ‘is it clear what we’re focusing on with this strategy?’ If the answer is yes, follow with ‘how will we achieve it?’.

The ideal answer begins with ‘we could...or…’.

Our strategies have worked best when they are broad enough to allow imaginative, independent ownership of decision-making by individuals or squads (and not telling them precisely what to do), but don’t make anyone confused about the direction. I find that a constant challenge.

Why is this different from having OKRs?

OKRs is a great framework to use alongside a clear strategy. At Lingumi, we use our strategy to frame and decide our goals. The key difference is OKRs are entirely unopinionated: ‘increase d30 retention by 5%’ doesn’t factor in any direction for how to do that. Strategies are opinionated: ‘Invest heavily in original content’ might be a Netflix strategy, with a retention key result attached, but it’s a very different strategy to ‘Strike major licensing deals on new content’, which might have the same key result and outcome. (By the way, if you are interested in Netflix’s strategy, Ben Thompson has covered it magnificently on Stratechery. TL;DR: it’s the former.)

How does this fit with key revenue or fundraising goals?

A friend, who is the CEO at a fast-growing post-Series A startup, asked me the following upon reading a draft of this piece:

The challenge I have with this is linking it to resources (time and cash). As well as long-term vision. E.g. how does what we're doing in this quarter, translate to a target for the year, which translates to a target for the next raise, which translates to a target for exit?

This is a tough question. I think the annual strategy should be quite pure, separated from the quarterly OKRs, revenue targets, fundraising and so on, and focus on fixing either a very broad but concrete customer problem (ideal!) or a distinct commercial problem. I say ‘pure’ in the sense that it doesn’t hint at a revenue target or fundraise, rather acting as a meta-narrative and decision-making tool for everything you’re working on during the year. Your teams can then use that to set their strategies, OKRs, and so on, and in my case, I’m now much less involved there.

However, I always write quarterly strategy slides, even if they’re very similar to the annual strategy. If, in a given quarter, the primary company problem is ‘we’re running out of cash but need revenue to be >£x to raise our next round’ then that quarter’s strategy might be quite different, or even the next two quarters’ strategy. We do sometimes deviate, but we do it rarely, deliberately, and acknowledge that deviation loudly in the team.

The danger here is when the quarterly strategy is distinctly divergent from the annual strategy - focusing on getting revenue to £x actively slows down or detracts from the annual strategy. If that’s the case, I’d question if it’s the right time to fundraise, the presumed requirements to raise are actually the necessary ones (I don’t think you actually have to have £1m ARR to raise a ‘Series A’), or if the annual strategy was the right one. I always say to the team that we are open to changing our strategy at any time, but it’d be a major decision, based on very clear evidence that the strategy is not serving us in pursuit of achieving our mission without running out of money.

Conversely, we recently made a slight reduction in a specific financial target in order to enable a trade-off in favour of our strategy: to hit our OKRs, we needed to make move B, but to follow our strategy, we had to make move A. We stuck with A, and reduced the target. I’d feel very confident explaining that choice to any investor questioning why we missed the goal. This is a marathon, not a sprint. Saying that, we have a fairly healthy balance sheet right now. I’ll test that confidence if and when that cushion is less comfy.

How does this approach to strategy develop as the company gets bigger?

What a good question, I’m wondering the same thing! If you’re writing strategies in this vein, but at larger team sizes, email me! First name at domain name dot com.

The simple answer is I don’t know. We’re already beginning to use strategies / maps / decks within teams (Tech, Product, etc) to convert the company strategy into sub-strategies, or departmental tactics, that feed back into the company strategy. Rooting it back to the core strategy means that the core bit has to be broad enough to encompass what a full team can achieve. At Lingumi, we’re a little under 40 people at the time of writing, and it’s working for us now, and worked when we were 8 people. I’ll update this or write a follow-up on how this method of communicating a strategy scales.

As mentioned above, if you’re a CEO at a similar stage, and want to spar more regularly on these topics, or join a selected peer group (from others who email me, if anyone does), let me know by email. First name at domain name dot com. I’ll aim to update this or add a second version of it in 6-12 months as I develop this strategy-writing process with further scale (fingers crossed).

If you have feedback on this piece, want to debate it, or do something similar, you can also send me a Tweet.

If you aren’t a subscriber to this newsletter but want my pieces in your inbox, subscribe here: