Gemini decision workflow: compare options without overthinking

A colleague once asked Gemini to compare three database technologies for a new project. Gemini produced a detailed feature comparison across twenty dimensions. My colleague spent two days reading it and still could not decide because most of those twenty dimensions were irrelevant to their specific situation.

I do it differently. Before mentioning any options, I write down my criteria: 'For this project I care about: learning curve (we have two junior developers), operational cost (we have a $500/month infrastructure budget), and community support (we need to find StackOverflow answers when things break). I do not care about features we will never use.'

Then I ask Gemini: 'Are these the right criteria for evaluating a database for a small team with limited budget building a content management system?' Sometimes Gemini suggests criteria I overlooked, like data migration tooling or backup automation. That pre-check makes the actual comparison much more useful.

The quality of your comparison depends entirely on how consistently you describe each option. If you write two paragraphs about option A but only one sentence about option B, Gemini will naturally favor option A in its analysis because that is where it has more information to work with.

For each option I include the same structure: what it is, how much it costs, what skill set it requires, what its main limitation is, and any specific context about my project that affects this option. Same format for all options. This forces fair comparison.

Be honest about the options, including the ones you are rooting for. If you already prefer PostgreSQL but put it on the list to 'validate' your choice, include its actual drawbacks: it uses more memory than SQLite for small datasets, and its JSON query syntax is less intuitive than MongoDB's for document-heavy workloads. Gemini is most useful when it challenges your assumptions, not when it confirms them.

Decision fatigue is real. Gemini produces its analysis when the information is fresh and the criteria are clear. Waiting a week to decide means you have lost the context, second-guessed your criteria, and possibly introduced new factors that complicate things unnecessarily.

I document the decision in our team wiki: which option, why, what tradeoffs we accepted, and what conditions would make us reconsider. This takes five minutes but saves the 'why did we choose this again?' conversation three months later.

Set a calendar reminder for three to six months out to revisit the decision. Did the chosen option perform as expected? Were our original criteria the right ones? This practice has improved my decision-making significantly because it closes the feedback loop that most decisions lack.