Analytics remains a top priority for CIOs and CEOs. Often, driving investment without clear evaluation of the resulting impact. As we've discussed augmented analytics including augmented data preparation, augmented data discovery, and augmented data science enables much faster creation of analytic content. As well, citizen roles such as the citizen data scientists, citizen developers, and citizen integrators are driving increased creation of analytic content through a broader engagement with the data. Data science and machine learning surveys often indicate that less than half of data science and machine learning analytic assets are deployed into production. Let alone, actually used. Sure, data science and AI remains speculative and experimental as they should be, and a lot of the work is thrown out. But even over the next few years, only 20 percent of analytic insights will actually deliver business outcomes. The growing availability of easy-to-use analytic tools has accelerated the creation of new analytic insights. What used to require SQL skills as we discussed earlier, is now a drag-and-drop self-service capability. The automation of data science at least partially has enabled the emergence of this new role of the citizen data scientists. This refers to any business user performing analysis that a few years ago would have required the skills of a full-time, highly experienced data scientist. Although the ability to create new insights can be a blessing for frustrated decision-makers, it exacerbates the potentially massive hangover of shelf work. This is analytic content that's been created but relegated to the shelf to gather dust instead of being incorporated into the business process to drive business value. This isn't a new phenomenon. There are ubiquitous anecdotes of reports failing to be updated for six months and nobody noticing, or data scientists who can't get operations to change a process because well, that's the way we've always done it. These episodes show that organizational use of analytic insight has never been close to 100 percent, and the number probably a significantly lower than many organizations realize. Now, this may not be a bad thing. Even just a single insight can have significant impact, delivering almost the entire return on investment for a project through a single transformational business concept. We've seen tons of examples of that earlier on. As the volume of analysis increases, the potential for that groundbreaking insight to be created also increases. The danger is that as analytics matures within organizations, the cost of staff and support analytic projects can no longer justify unrealized experimentation. Teams that aren't well organized or consistently operationalizing their findings and focusing on business outcomes will struggle to justify additional investment, additional spending, and additional hiring. In the case of large shelf work investments with capitalized labor component to them, it can be even more detrimental with created assets needing to be impaired or even written off, leading to unplanned financial accounting burdens.