layout: true class: middle, center <header> <p class="left"></p> <p class="right"></p> </header> <footer> <p class="left">@CraigBuchek</p> </footer> --- class: title, middle, center # Too Late to Estimate ### Craig Buchek ## Hofstadter’s Law: It always takes longer than you expect, even when you take into account Hofstadter’s Law ??? * I'd like to start by asking some questions * Please yell out your answers --- * Are you pretty good at estimating? * Or pretty bad? --- * Do you usually estimate too high? * Or too low? --- * Why do we estimate? ??? * Why are we asked for estimates? * Answer: * To know how much something will cost * Other potential answers: * To know when we'll be done * To know if we're on schedule --- * Why do we want to know how much things will cost? ??? * Why do we want to know how much things will cost? * Answer: * To know if it will be worth the cost --- * How do we know whether it will be worth the cost? ??? * How do we know whether something will be worth the cost? * Answer: * Compare the cost with the value * Compute the ROI --- * ROI - Return on Investment ~~~ value - cost value ROI = -------------- = ------- - 1 cost cost ~~~ ??? * ROI is the formula we use ... * To know whether something is worth the cost --- * What do we want our ROI to be? ~~~ value - cost value ROI = -------------- = ------- - 1 cost cost ~~~ ??? * What do we want our ROI to be? * Answers: * Positive * Higher than the ROI of other alternatives --- * How do we know whether we should start something? ??? * How do we know whether we should start something? * Answers: * Has to have a positive ROI * Has to have higher ROI than alternatives --- * Can we calculate ROI without knowing the expected value? ~~~ value - cost value ROI = -------------- = ------- - 1 cost cost ~~~ ??? * If we DON'T know the value of a project ... * Can we figure out the ROI? * Answer: * NO! --- * Do project managers typically determine value before asking for cost/time estimates? ??? * Do project managers usually ... * Determine value before asking for estimates? * Answer: * No --- * Conclusions * We're worse at calculating ROI than estimating * No point calculating ROI once we've started * Don't estimate at story level ??? * Once you've started the project ... * Too late to determine ROI * Too late to estimate at fine-grained level --- * \#NoEstimates ??? * So if we don't estimate at the story level, what can we do? --- * Estimate only at project or epic level ??? * \#NoEstimates doesn't necessarily mean **NO** estimates * I use it to mean no **fine-grained** estimates * Estimating at higher levels is still OK * Maybe even necessary --- * Shorten the feedback loop ??? * Or we can minimize losses * Assess **actual** ROI after each iteration * Stop when ROI is no longer worth it --- * Count stories instead of story points ??? * Number of stories provides better predictive value ... * Than story points * Adding points to stories is a waste of time ---  * @CraigBuchek ??? * I've got a blog article * Problems with estimation * Or tweet me