Train the monkey before building the pedestal
When we’re developing new products, platforms or services it’s important that we have a guiding light – a north star that we can use to help us keep heading in the right direction, and avoid going off track when we start to get technology involved.
Astro Teller, Google X’s Captain of Moonshots, says: “If you want a monkey to juggle flaming torches while standing on a pedestal in the town square, start by seeing whether you can train a single monkey to juggle the torches. Building the pedestal is the easy part; it’s pointless if the monkey-juggling turns out to be impossible.”
This sums it up perfectly: people buy into the vision because it would be amazing to see, so budgets are allocated and promises made, Reputations are now resting on the project and so people want to show progress quickly. So, a flurry of tasks are picked off the ‘easy’ pile – invest in the technology, get it installed and running, demo its capabilities and get some ‘wow’ responses. But these are the pedestals.
If the monkey can’t juggle, then you end up with a marvellous pedestal, but the grand idea that started the whole project is nowhere to be seen. The engineers are proud of the technology they’ve built (and rightly so), but ultimately the success they were measured against is pointless if the rest of the project proves fruitless.
Zero time has been invested in understanding whether the monkeys could actually be trained, how long it would take, and how useful that would be. Everyone has simply worked on the assumption that it was a good idea, because they were told it was.
In the real world
Having grand ideas is great – but we need to proceed from them being 95% hope and 5% reality to a stage where we know the idea will work, before we start allocating large budgets and reorganising teams. To come back to Astro’s quote, we need to somehow test the hypothesis that a monkey can juggle fire (and that this would be helpful) before we build the pedestal.
Following the straightforward mantra of ‘if you have to fail, fail fast, but more importantly fail cheap’, project teams should focus on designing the smallest set of experiments possible to show there’s value in the idea. Rather than only looking for reasons why it’s a great idea, they must look to try and disprove it too, and heed warnings that it might end up being a total waste of time and money.
A series of hard tests should tear the idea to shreds, and only if it passes all of these tests should the pedestal building begin with intent and purpose. Indeed, by undertaking this process, not only can teams shape the idea more firmly, they may find that the pedestals are unnecessary – people will come to watch the monkey juggle whether it’s elevated or not!
A project example
When a company announced their new implementation of AI had been successful, a competitor wanted to replicate this in their own business. The prize was grand – they could show thought leadership and innovation in their industry, and they could also save money and increase customer satisfaction.
They allocated a multi-million pound budget, put together a project team, and gave the green light to replicate the AI tech being built for them.
But they missed the importance of ‘if’. If the technology and process could give accurate predictions based on the data that they could collect. In this way, they confused the output with the process. They saw the competitor’s success and assumed the process that led to it was the visible parts (the flashy AI tool, the announcements and the kudos).
An assessment at the start showed they needed to design a set of experiments, which would reveal what data they required and how accurate they could be. A year and a half later, a further assessment looked at the project outcomes.
This second review revealed they had a lovely looking user interface but the tool was actually costing their teams more than it was saving, because of its false positive rate (they were spending more time investigating misclassified problems than actual problems). It turned out none of the experiments proposed in the initial assessment had been done, with no early validation of feasibility completed.
Instead, the company wanted to demonstrate progress – so they began to build more features in their app and simply hoped the AI could be patched in when ready. But when this AI never materialised, the whole project was canned and they went back to using a spreadsheet.
The solution
It’s entirely understandable that companies want to focus on the prize – following the direction of the north star. But it’s important not to be blinded by it, to seek objections and blocks before committing to building the technology which underpins the project.
The crucial experimental phase at the beginning of the project might not feel like the sort of exciting progress everyone is looking forward to seeing – but by starting slowly and deliberately, this allows confidence and momentum to build. Ultimately then, by the time the tech building begins, the end result should be virtually guaranteed to succeed.
No empty pedestal left to stand unused, a reminder of money and time spent needlessly because everyone was too eager to skip to the ‘exciting’ part that they neglected to test the idea properly first.