AS3DM or Agile Scientific Data-Driven Decision Making for the win! but what is it?
Only 32 percent of business executives say they create measurable value from data, and just 27 percent report that their data and analytics projects produce actionable insights*.
As a former data scientist and a data science manager, this concerns me.
I strongly believe in the power of data and the value of data science expertise. So, why do so many ‘data-driven’ initiatives fail?
The answer is that being data-driven or even AI-driven is not enough.
Data can be misinterpreted or manipulated, and automation doesn’t solve that.
That’s why I advocate for Agile Scientific Data-Driven Decision Making (𝐀𝐒3𝐃𝐌).
➡ Here are some of the key practical components of AS3DM compared to typical Data-Driven Decision Making.
𝐅𝐞𝐞𝐝𝐛𝐚𝐜𝐤 𝐌𝐞𝐜𝐡𝐚𝐧𝐢𝐬𝐦𝐬: AS3DM establishes feedback loops for continuous improvement, whereas typical data-driven decision-making (DDDM) lacks ongoing feedback mechanisms, often prioritizing marginal wins over continuous improvement.
𝐋𝐨𝐧𝐠𝐢𝐭𝐮𝐝𝐢𝐧𝐚𝐥 𝐒𝐭𝐮𝐝𝐢𝐞𝐬: AS3DM considers longitudinal studies and long-term trends, in contrast to DDDM, which often focuses only on short-term data and misses longer-term insights.
𝐁𝐚𝐥𝐚𝐧𝐜𝐞𝐝 𝐔𝐬𝐞 𝐨𝐟 𝐀𝐮𝐭𝐨𝐦𝐚𝐭𝐢𝐨𝐧 𝐚𝐧𝐝 𝐄𝐱𝐩𝐞𝐫𝐭 𝐈𝐧𝐭𝐞𝐫𝐯𝐞𝐧𝐭𝐢𝐨𝐧: AS3DM balances automation with expert intervention to ensure insights are both accurate and contextually relevant. In contrast, DDDM either over-relies on automation, missing nuanced contextual factors, or under-relies on it, leading to wasted efforts and resources.
𝐏𝐞𝐞𝐫 𝐑𝐞𝐯𝐢𝐞𝐰: AS3DM incorporates peer review to ensure accuracy and reliability, whereas DDDM often involves decision-making in silos without peer evaluation.
𝐂𝐨𝐧𝐭𝐫𝐨𝐥𝐥𝐞𝐝 𝐄𝐱𝐩𝐞𝐫𝐢𝐦𝐞𝐧𝐭𝐬: AS3DM conducts controlled experiments to isolate specific variables and measure their impact for clear insights, while DDDM often emphasizes data volume over analysis rigor.
𝐒𝐜𝐢𝐞𝐧𝐭𝐢𝐟𝐢𝐜 𝐀𝐩𝐩𝐫𝐨𝐚𝐜𝐡𝐞𝐬: AS3DM employs scientific methods such as hypothesis testing to guide analysis, in contrast to DDDM which is prone to the data-dredging or p-hacking pitfall.
𝐌𝐢𝐭𝐢𝐠𝐚𝐭𝐢𝐨𝐧 𝐨𝐟 𝐁𝐢𝐚𝐬𝐞𝐬: AS3DM involves structured processes that mitigate biases, whether they are cognitive, statistical, or agenda-driven unlike DDDM, which often lacks such clear bias-mitigating processes.
𝐀𝐠𝐢𝐥𝐢𝐭𝐲 𝐰𝐢𝐭𝐡 𝐓𝐫𝐚𝐢𝐧𝐢𝐧𝐠 𝐚𝐧𝐝 𝐓𝐨𝐨𝐥𝐬: AS3DM is agile in selecting appropriate technological tools and ensures proper training for their use, while DDDM continues using outdated or suboptimal tools due to technical debt or biases.
𝐊𝐧𝐨𝐰𝐥𝐞𝐝𝐠𝐞 𝐒𝐡𝐚𝐫𝐢𝐧𝐠: By promoting the sharing of ongoing projects and methodologies, AS3DM helps avoid redundant efforts. This approach contrasts with typical DDDM practices, where teams often operate in isolation, leading to duplicated efforts.
