The Advocacy Iceberg

The Advocacy Iceberg

October 2, 2013

Every hand’s a winner and every hand’s a loser

The Advocacy Iceberg
The Advocacy Iceberg
Every hand's a winner and every hand's a loser
Subscribe on

You have no subscribe urls set, please go to Podcast → Settings → Feed Details to set you your subscribe urls.

Show Notes


The kid spread his hand and then began to blush

But his face turned pale when he saw my queen-high flush


T Bone Walker there, describing how to play a winning hand of poker.

Though in fact (a) he’s not actually talking about poker and (b) it’s rarely that simple.

According to Nate Silver, in his book The Signal & The Noise, a good poker player can still be financially behind after tens of thousands of hands, if the cards go against you, at the unluckiest end of the spectrum.

Good players are highly skilled, but it’s also about how the cards fall.

So poker is a high skill/high luck game. Nate Silver sets it out like this:



Low luck

High luck

Low skill



High skill




This seems to me to be a helpful way to think about campaigning too, which I would put firmly in the bottom quadrant.

There’s good campaigning and less good campaigning, and a fair amount of consensus about the difference (though there would be areas of disagreement).

But a good plan well executed can fail, and a bad plan poorly delivered can succeed. There’s no linear correlation between skill in delivery and result. And that’s because of the different influences that come into play along the way.

As with poker (and for complex systems generally) some of the key characteristics in campaigning include that:

  • There is high sensitivity to initial conditions – with events then potentially following radically different paths, with continual forks in the road along the way
  • Results are relational, outcomes are generated through interaction rather than being determined by one player
  • Feedback (including the response of others to events and information) can either promote or inhibit change
  • Results can be disproportionate and often difficult to predict
  • Results occur within a historical system, how things have gone before influence the nature of future trajectories
  • Low probability events may be rare but they do occur (as T Bone Walker identifies)

I’m not sure this comes down to ‘luck’ – this has connotations of a kind of passive good fortune, whereas in campaigning you often make your own luck (but then the same is true in poker).

But the point is to highlight the role played by external forces outside of the campaign’s control. And whilst campaigners Intellectually understand and accept this way of thinking, it’s not fully followed through in practice.

Some suggested implications are that:


1.     Campaigns can beat the odds

The absence of a linear read-across from inputs to results is what can make campaigning so potent.

However, it also makes campaigning as a discipline sometimes difficult to shoehorn into organisational ways of thinking, which tend to be more comfortable with more predictable and reliable formulas.


2.     Think probabilistically in planning.

Traditional planning techniques may help foster an illusion of certainty, and of predictability. Whereas the future is likely to be intrinsically uncertain, even if we have enough information to make well-founded judgements about it (which often we don’t).

So it’s about finding ways to integrate this strategic uncertainty within planning processes, rather than ignoring it.

Decision trees could be one tool for this, and are an under-explored area in campaign planning I think.

Plus it seems like there could be good possibility of deploying techniques based on Bayesian probability (also covered in Nate Silver’s book, and which I might write more about shortly).


3.     Doing as planning

A complementary approach is to place less overall emphasis on anticipating the future and more on adapting to the present. To shift from ‘predict and control’ to ‘measure and react’, continually reassessing prospects along the way.

This puts a high premium on being able to read the game and understand the odds, and also implies that the distinction between ‘planning’ and ‘monitoring’ should be a bit more blurred.


4.     Be explicit about the role of external factors when conducting evaluations and reviews.

In one recent campaign evaluation, for example, we ran an exercise with staff, asking them to plot the campaign on two dimensions

  • how effectively the campaign was delivered, and
  • the extent to which it benefited from good fortune, or was hampered by ill fortune

Using a model like this:



+ ive












– ive




– ive                                                                                                          + ive






This seemed like a valuable exercise in that external factors are typically acknowledged as important but are often not so visible.

Following on from this way of thinking, a simple representation of a campaign would ideally capture 3 things:

  • significance of the outcome
  • importance of the organisation’s contribution
  • extent to which external factors were conducive to success


5.     Assessing the results of single campaigns in isolation may give an unreliable picture

The suggestion that funders, for example, should adopt a portfolio approach – looking across the programme rather than expecting results every time – in assessing advocacy is helpful here.

These same principles need to be translated to, and picked up in, single-organisational contexts.


6.     Know when to walk away and know when to run

If you are a good player, it will pay off in the longer term. Though it becomes difficult to be sure after a time of losing whether you are a good player suffering bad luck, or just not such a great player in the first place.

On the other hand, winning may create an illusion of high effectiveness.

So it’s not just about the results, it’s about building the best strategic understanding you can of the underlying factors driving the results, and not being too self-centred about what the reasons might be.