The Marines Warfighting Doctrine
I often turn to military metaphors when I’m trying to explain management concepts.
The reason is simple: the military have been doing this a lot longer than most industries and they operate in a far more chaotic, critical and life threatening environment than most enterprises.
Most management translations of military theory are dead and bleached versions of some Sun-Tsu aphorism, quoted out of context and with dubious relevance.
So I thought I’d try and go back to some original sources and see if the metaphors made sense in a modern management context.
One of the most common issues I see in business is the disconnect between strategy and it’s execution. In fact many executives can’t even distinguish strategy from it’s execution.
Here’s what the Marine’s war-fighting manual has to say on the subject of strategy and tactics:
The highest level is the strategic level… Strategy involves establishing goals, assigning forces, providing assets, and imposing conditions on the use of force in theaters of war. Strategy derived from political and policy objectives must be clearly understood to be the sole authoritative basis for all operations.
Translation : Strategy involves establishing goals, assigning resources and setting rules. Every activity must contribute to achieving the strategy, which must in turn support the reason for an organisation’s existence.
The lowest level is the tactical level… The tactical level also includes the technical application of combat power, which consists of those techniques and procedures for accomplishing specific tasks within a tactical action. These in- clude the call for fire, techniques of fire, the operation of weap- ons and equipment, and tactical movement techniques. There is a certain overlap between tactics and techniques. We make the point only to draw the distinction between tactics, which re- quires judgment and creativity, and techniques and procedures, which generally involves repetitive routine.
Translation : Day to day activity consists of accomplishing specific tasks using techniques, procedures and tools. Some tasks require judgement and creativity and other tasks require repetitive routine.
The operational level of war links the strategic and tacical levels. It is the use of tactical results to attain strategic objectives. The operational level includes deciding when, where, and under what conditions to engage the enemy in bat- tle—and when, where, and under what conditions to refuse bat- tle in support of higher aims. Actions at this level imply a broader dimension of time and space than actions at the tactical level.
Translation: To link strategy with day-to-day activities requires organisation. It requires choices to be made about what activities to engage in to achieve strategic objectives and what tools and techniques should be deployed in what situation.
Pretty good so far.
How about decision making? Another key weakness in business.
Time is a critical factor in effective decisionmaking—often the most important factor. A key part of effective decisionmak- ing is realizing how much decision time is available and mak- ing the most of that time. In general, whoever can make and implement decisions consistently faster gains a tremendous, of- ten decisive advantage. Decisionmaking in execution thus be- comes a time-competitive process, and timeliness of decisions becomes essential to generating tempo. Timely decisions de- mand rapid thinking with consideration limited to essential fac- tors. In such situations, we should spare no effort to accelerate our decisionmaking ability. That said, we should also recognize those situations in which time is not a limiting factor—such as deliberate planning situations—and should not rush our deci- sions unnecessarily.
We should base our decisions on awareness rather than on mechanical habit. That is, we act on a keen appreciation for the essential factors that make each situation unique instead of from conditioned response. We must have the moral courage to make tough decisions in the face of uncertainty—and to accept full responsibility for those decisions—when the natural incli- nation would be to postpone the decision pending more com- plete information. To delay action in an emergency because of incomplete information shows a lack of moral courage. We do not want to make rash decisions, but we must not squander op- portunities while trying to gain more information.
Finally, since all decisions must be made in the face of un- certainty and since every situation is unique, there is no perfect solution to any battlefield problem. Therefore, we should not agonize over one. The essence of the problem is to select a promising course of action with an acceptable degree of risk and to do it more quickly than our foe. In this respect, “a good plan violently executed now is better than a perfect plan exe- cuted next week.”
Couldn’t have said it better myself.
But surely the military situation is different? Military structures and discipline allow for a level of command and control that is impossible in a fluid, commercial situation… right?
Anyone who believes that has no concept of what war fighting is actually like.
Autonomy and initiative are essential to military operations and flexible fighting forces like the Marines understand it better than any ‘agile’ commercial organisation:
There are two parts to any mission: the task to be accom- plished and the reason or intent behind it. The intent is thus a part of every mission. The task describes the action to be taken while the intent describes the purpose of the action. The task denotes what is to be done, and sometimes when and where; the intent explains why. Of the two, the intent is predominant. While a situation may change, making the task obsolete, the in- tent is more lasting and continues to guide our actions. Under- standing the intent of our commander allows us to exercise initiative in harmony with the commander’s desires.
The military may not be a perfect mirror for modern commercial enterprises, but by ignoring the lessons of their history, you will be doomed to repeat them.
Confusion Reigns (Z+15)
Sailing Primer Campaign – day 4
- Expenditure = $133.87
- Impressions = 34,711
- Clicks = 160
- Decisions = many, many on the fly
- Progress = questionable
- Sales = 0
- Confusion = rising
Well, we’re not millionaires yet.
In fact we’re not even profitable.
To be brutally honest, we haven’t made a red cent – not one.
The table above tells the story pretty accurately. We’ve had the campaign for the sailing primer up and running since Friday last week and 34,000 people have ‘seen’ our advert. Out of those, only about 160 have clicked on it (0.46% Click-Through-Rate or CTR) and a solid total of 0 have bought the Primer for $7.95.
Not stunning.
On day one we launched it with 3 groupings from 11 keywords. Yesterday we made some changes to try and tweak the performance of each keyword. Today we made some more.
The decision for each keyword is simple binary : keep it or lose it. Beyond that there are a number of variables you can play with including the Max CPC and the text of the keyword. Beyond that are a number of options you can play including “keyword match type”, “negative keywords” and “keyword” rotation. More on those later.
Choosing what to keep and what to lose (and the other options) is supported by a plethora of data, all of which is confusing and contradictory.
My own private theory is we’re in the zone of chaos (see the Cynefin decision model) and we shouldn’t spend too long analysing the thin data that we’ve garnered. We need to make decisions, pose hypotheses and test them out (PDCA). If they’re borne out by the subsequent data then they stand to become the basis of our model going forward, if not they’re discarded.
But one simple way I found to make use of this data is to use a “quartile heatmap” – that is, assume the data follows a normal distribution and map which data items are in the upper quartile and the lower quartile of a normal distribution.
The following screenshot shows the data for our keywords in this heatmap. Green indicates the ‘upper’ quartile and red the ‘lower’ quartile.
Quartile Heatmap for Keywords
For example some keywords that have high clicks and high click-through – we obviously want to keep those. But there are others, like “How to Sail”, which have a relatively low CTR but high clicks – what to do with them?
In this instance we decided to keep them because we need the data. If we eliminated them on the basis of CTR we’d lose valuable clicks. Once we start getting sales and tuning the rest of the pipeline, we could theoretically eliminate them in favour of a higher CTR keyword.
And so on… and so on…
We are conscious that, at the moment, we might be sub-optimising our pipeline. Since we don’t have sales, we can’t draw end-to-end conclusions about the pipeline and are forced to make data based on interim goals, like driving traffic and driving impressions.
C’est la vie.
A learning along the way is that Adwords is not user friendly. Google goes out of it’s way to make it helpful, in the same way that owning a quick-start guide for the Space Shuttle would help you fly it.
Google automates a lot of the initial choices for you in Adwords but, coincidentally, they do it in a way that would seem to maximise their own revenue. Quelle suprise!
For example, we wanted to run multiple adverts in parallel to test which is the most effective. So we set up three ads in parallel and ran them for a couple of days.
Adword Distribution in a Campaign
When we came to check the ads we discovered that #1 had been shown 82% of the time, #2 only 12% of the time and #3 a paltry 6%.
It turns out there’s a setting buried in the “Advanced Settings” screen for your Campaign which chooses to optimise ad display for clicks, for conversions (if you have conversion code) or not at all – which is not recommended!
But that’s only applicable if you use the “advanced campaign” setting from the very beginning. If you use the “standard” campaign setting, you never see any of this and Google choose to optimises your ads by running the one that is likely to get the most clicks. This automatically deprecates some ads in favour of one that gets clicks but doesn’t tell you which one leads to more sales – in effect optimising revenue for Google ahead of useful data for you.
There’s a similar setting in Campaign Settings for Bidding and Budget which automates the max bid for a keyword, or not.
The blurb in the help says :
If you’re new to AdWords (or if you’re busy), we recommend this option. All you have to do is set a daily budget, and the AdWords system manages your bids for you, to bring you the most clicks possible within your budget.
In essence, Adwords will raise your bid (if necessary) in an attempt to spend all of your daily budget. It may well bring you the most clicks in a day, but it won’t do so econmically or predictably and, once again, the default option maximises revenue for Google.
Now if we can only find the option marked “maximise revenue for vendor” and turn it on 😉
3 rules for new founders
Lean is good for companies but bad for people. It doesn’t help with that potentially lengthy period when you’re still trying to find an idea with legs and don’t yet have funding or revenue. Here’s how to handle that bit.
The Start of Week Two (Z+8)
I had my first official dummy spit yesterday.
Another theoretical argument at the end of a long day of theory pushed me past my breaking point. I lost my bottle and stormed out, declaring I’d had enough. Enough of what I wasn’t sure but I did know that if I had stayed in the room, someone else was going to suffer.
There’s a tension between theory and practice in any pursuit.
I have often been accused of erring on the side of the philosophical and of being too academic and not action oriented. I usually countered this with : “Would you like me to understand what it is that I’m doing, or should I just make it up as I go along?” – implying of course that my accuser is making it up, as they often were.
In MUSE we’ve established the principle that our decisions should be based on cold hard data, not hunches.
This entails a certain amount of discussion and a certain amount of confrontation – we openly challenge weak language or weak thinking on the part of another. On the whole it’s done us proud by limiting our assumptions and keeping us safe from the pitfalls of a blind assertion.
But yesterday I was feeling the call to action and the frustration at not being able to settle the theoretical question boiled over inside of me.
In reflection, it wasn’t the theoretical debate that really frustrated me. It was the lack of structure and progress in our deliberation that niggled me. It was late in the day, we were all unfocussed and no one including myself could untangle a path out of our circular discussion. We went round and round and we all felt the need to continue. To make progress. To achieve something.
In 20:20 hindsight, the decision to stop was a good one, but the execution lacked a little flair as I’m sure my partners would confirm.
We recovered today with some solid progress and some more (fruitful!) theoretical discussions.
Overall, the week has started well, with the decision to adopt a weekly “no-fail goal” – a shared goal which takes precedence over everything else – paying off handsomely. The idea (from the Pollenizer startup book) is to focus the team on a stretch goal for the week that drives your collective effort towards a single end.
Our was easy, and audacious : we want revenue in the bank this week.
It was amazing how fast this cut through our thinking.
Of our two candidates, only one has the potential to generate instant income. It’s a digital information product which requires no manufacture, no storage and no physical distribution. We can ship product and take money with only an Adwords campaign, a Google Site and a Paypal account.
This has the bonus of allowing us to build a fully functional model of the first half of our supply chain, without the complications of suppliers, warehouses or fulfilment. It allows us to tinker with half the variables in the equation without worrying about the other half.
So far we’ve managed to brainstorm a campaign, set it up in Adwords, setup the DNS, skeleton website and Paypal account and start filling out the skeleton with actual content. We still have questions over our e-fulfilment but in true Lean style, we’ll manually send out the files until we have built some minimal automation.
And just in case you think we’ve been idle, we did manage a couple of other things this week.
On Monday, we spent half a day in the Hay St mall validating our other surviving product/candidate. We thought it important to get out of the office and make contact with reality so we trooped into town and spent two hours bailing up total strangers and subjecting them to a barrage of questions. Results were encouraging enough for us to not-kill the candidate and it remains on the board, awaiting its next step.
We also appointed an accountant to setup a company for us. We spent an hour with him investigating the finer points of small company operation and think we’ve found a man we can trust. We’re going to engage his firm to handle our tax and company matters for us and setup MUSE under the umbrella of a small proprietary company.
The choice of company is important for us, our venture is inherently risky and we have assets to protect. It is slightly more costly than a partnership but we think it offers us much more flexibility in the long run.
More on that later….
End of Week 1 (Z+4)
The first week draws to a close.
I knew this was going to be a roller-coaster ride but I didn’t expect to hit all the highs and lows in the first week! The euphoria of Day Zero was quickly overtaken by the grind of hump day (Z+2) and a second wind that got us moving again on Thursday.
The pace has been frantic already and emotions have been battered by the winds of change.
Yesterday we confronted the lurking issue of burnout, brought on by the pace and the stress of a new environment. We all feel an obligation to each other that drives us to try and push our progress along, however we can.
But there’s a significant difference between being busy and being productive, and there’s a fine line between a healthy sense of duty and an unhealthy burden of guilt. We should be doing things because it drives the business, not because we are worried about what someone will think.
So, today we’ve granted ourselves a slow day, working from home, looking after our other clients and taking some downtime to reflect.
I think we’ve made good progress this week.
We’ve made six go/no-go decisions about potential products and we have three two viable candidates on the board, each awaiting the next test. The latest of these is a pure information product that will carry no production cost and no shipping cost. We debated whether we should push this one ahead since it doesn’t touch the full product lifecycle but we decided that the positives outweigh the negatives.
We can test the first third of our process getting adwords live, the website, a shopping cart, payment gateway and delivery system. It will also give us live contact with real customers. This will give us the ability to A/B test, to learn the vagaries of Google AdWords and to validate our thinking for almost zero cost.
It also gives us a way to evaluate some of our decision criteria by eliminating production, shipping and handling costs, but leaving cost of acquisition to compare to revenue.
And finally it might put some cash in the bank.
Day Zero
The Card Wall
Our first day as a startup turned up a lot of surprises.
The hard grind of analytical/emotional arguments wasn’t a surprise. Neither were the false starts, the pauses for clarification or the finger waving in each other’s faces. We’d spent a couple of long months grinding the edges off our interactions to ensure that we could live with each other and hold an argument without leaving blood on the carpet.
One thing that surprised me was how quickly we were able to strip away the peripheral things – the soulless cement walls, the lack of air conditioning and the uncomfortable chairs – and focus on the brutal facts. Within 30 minutes we had our first candidate running the gauntlet and within an hour it had fallen. Two more were to follow within the next hour.
Onno was startled by the fact that he grokked everything – that none of us fell behind as we stampeded through our first candidates. Despite all the prep, this was new territory for us and fear and uncertainty loomed large as we ventured into the unknown.
The biggest surprise though, was that we couldn’t kill all of the candidates.
Despite our best efforts, one survived.
Day one and we have a (partially) viable product – who’d have thunk it?
The unlikely candidate has yet to run the full gauntlet, but it survived the first round and it’s still standing. Three other candidates have gone by the wayside, perishing in the fiery crucible of the pitch: no viable market; no value proposition; not enough markup; no supplier.
But, unaccountably, this one survived.
Once we got over the shock, the debate was “do we stick to the formula or do we abandon our carefully crafted process?” The answer was easy – we need to validate. We need to go to the gemba and test our assumptions in reality. We crafted a market test that would let us validate our survivor with minimal effort and continue our process with confidence – homework for tonight.
The other debate that this produced was, “do we keep running with our survivor or do we continue to sanity test the other 129 remaining candidates?” We decided to split the difference and opt to continue testing one candidate a day, so we could keep learning, while we spent the rest of the time pushing our survivor through the gauntlet.
Tomorrow sees us taking the next step, let’s see how we maintain the pace….
Blogs I Follow
muse_startup 