Launching a new idea? First get YODA

Your Own DAta (YODA) is market data collected firsthand by your own team to validate your own idea.

I came across YODA in Alberto Savoia’s brilliant book, the right it: why so many ideas fail and how to make sure yours succeed. The key thesis is that we make very big business decisions based on data we find in the public domain: a research by a reputed consultancy, an insightful infographic from a leading newspaper, a scholarly hypothesis based on interviews by reputed professors for their book. These data points that we gather shape our broad strategic direction: which segment to target, which geography to apportion more of our budgets to. These strategic calls are made by very senior people in very serious settings, usually boardrooms.

I worked for a decade in the newspaper industry and let me give you a possible situation that may be playing out. Someone like me, I was responsible for strategy, who keeps abreast of all the trends, sees a series of paid newsletter launches by competition and also by individual journalists on platforms like Substack. I come up with a plan to launch paid newsletters written by some of our well known journalists. We know that the news industry has finally accepted that reader revenues will be more important than advertising dollars going forward. My newsletter plan fits the moment and I build a case showing how the top writers on Substack have racked by run rates of six figures in no time. Everyone gets excited, the board loves my idea, the journalists love it (I build in a small revenue share with them) and we are ready to go.

I get to put together a small product team and we decide to build the technology ourselves, so that we can tweak it as we want and build in various measurement tools suited to our needs. An off the shelf product won’t work. We do the maths and the cost of building the platform with all the social and payment integrations we want comes to half a million dollars. I take it to the board and it’s approved, the cost-benefit analysis is most convincing (I am very good with numbers).

We get building and after six months (in record breaking development time, we tell everyone) we launch five newsletters on subject matters based on the journalists who want to have a go first. There has been a bit of a budget overrun (we wanted to speed up the development, you see) and we ended up spending a million dollars.

In six months we know that the numbers are not even half of what we had projected. Of the five newsletters, only one has some traction. The writers of the other four have lost interest and editorial want to scrap the whole plan. I try and push it for a few more months. By now, I have spent a quarter of a million dollars on marketing the newsletters. We finally decide that not enough people are going to pay and we decide to go free and the newsletters’ role is now to help in retention and the content is handed over to trainee reporters.

Does this sound familiar?

How could this have been avoided?

The root of the problem was, as you would have caught on already, we took data available in the public domain from multiple sources as gospel and built our own castle on it. The first important fact to keep in mind is that everyone has their own motive. A world renowned consultancy writes about the changes wrought by the pandemic and highlights how work-from-home has IT challenges that needs a five point action plan. The consultant is writing thought leadership papers for one reason, to get projects.

A newspaper writes a piece based on a research with ‘30 top CEOs’ which concludes that ‘majority of CEOs think we are social animals, and we will be back in offices as soon as we can’. If you delve enough, most of us don’t, you may find the research has been commissioned by a large commercial property player who also coincidentally has taken a full page ad in the newspaper launching a swanky new office building.

Even with the cleanest motives, a professor writing a piece with a theory in his head tends to pick (almost subs consciously, we do it too) the most convincing data and ignores the red flags as outliers.

Which is why we need YODA’s wisdom. In his book, Mr Savoia states the following criteria needed for data:

1. Freshness: the data has to be fresh, as recent as possible. More than a few months may render the data obsolete depending on the category. There have been a number of launches in the packaged food industry in the recent months in India. CEOs have spoken of this as an area of sure growth, basis the huge uptake during the pandemic. Do people like to eat packaged food by choice? Convenience is a big driver, my guess (any recent research?). What will happen when all the cooks are back? Fresh food vs packaged food. Hmm. Do we have some data from this week, now that the vaccinations are going up?

2. Strong relevance: the data must be directly relevant to the problem at hand. Seems obvious. But. The fact that a lot of people throw away the cut onion packets that come with ordered-in kebabs or other starters in India doesn’t mean you shouldn’t have onions in that fantastic Greek salad you are planning to launch which will use only organic produce.

3. Known provenance: This we have already read about earlier. You cannot rely on data from sources whose underlying motivations are unknown.

4. Statistical significance: Again obvious but how often have we made investments in products based on a senior executive’s personal anecdotal data points. ‘All my 8-year old’s friends prefer white pasta as a snack nowadays, chips are so out’. The data must be from a statistically large group and collected based on a scientific design.

To meet these criteria, you have to get your own data. Other people’s data is just not good enough.

This will lead to better hypotheses (or guesses). You still have to test the idea but by spending a million and a quarter? Is there an alternative?

Yes, there is. But that’s a story for another day.

Have a good weekend.

Suprio

---

I am a partner at AcceleroBiz LLP. We deliver growth ideas, tested for market fit, at speed.