Is my prescription ready?\nHas my insurance claim been approved?\nHas my gate changed?\nThese are questions industry analysts at Forrester use to illustrate what they call the \u201cmobile mind shift,\u201d which they define as an expectation that \u201cI can get what I want, in my immediate context and moments of need\u201d on my smartphone.\nThe reflexive way people reach for their smartphones nearly anywhere, anytime, may even call to mind Pavlov\u2019s dogs.\nIvan Pavlov won a Nobel Prize\u00a0for describing how animals (or humans) can be trained to respond in a certain way to a particular stimulus. In the dogs\u2019 case, he observed dogs in the lab drooling when they saw someone in a white coat \u2014 because, as he figured out, they were always fed by someone wearing one. He famously went on to train the dogs to salivate when he rang a bell.\nAs executives, strategists or implementers within an enterprise, we may be tempted\u00a0to imagine that we are in the role of scientists like him, conditioning consumers to instinctively turn to our products and services during their mobile moments.\u00a0But it is more likely we are the ones who need new to learn new reflexes and build new muscle memories for how we handle investing in technology.\nIn short, when it comes to digital transformation, we\u2019re not Pavlov. We\u2019re the dogs.\nOur instinct to \u201cplan, discuss and control\u201d needs to be replaced with \u201cdoing, testing and discovering.\u201d This may feel uncomfortable because our organizations have trained us to associate safety with decision funnels, net present value (NPV) spreadsheets and final-approved plans in PowerPoint (complete with voluminous appendices of supporting data).\nWhat I would describe as the \u201ctraditional model\u201d for investing in technology reflects the limitations of its time:\nWe must winnow down our options into a static plan for releasing large amounts of code, months apart, by internally debating the meaning of extrinsically collected data.\nBut mobile, cloud and machine learning have changed the balance of risk and reward.\nAmazon\u2019s CEO pioneered a no-holds barred embrace of these implications with an \u201cAPIs everywhere\u201d mandate in 2002. (This presentation by an AWS devops manager on how they operate today is a must-read.) Today, the technological tools needed to take up continuous integration and relentless, hypothesis-driven (and often automatable) testing \u2014 not to mention the ability to immediately roll back deployments when necessary \u2014 are now mainstream and widely available to any company.\u00a0(Disclosure: my employer, Apigee, provides a platform for using APIs at scale and as strategic assets.)\nIn this context, the old way of doing things must give way to a new model that takes full advantage of this:\nWe can release small amounts of code, weeks apart, with ways to test our hypotheses for \u201cwhat\u2019s right\u201d and \u201cwhat\u2019s next\u201d defined in advance \u2014 and often built-in.\nHowever, individual behavior and organizational incentives must adapt in order to make the most of it:\nIf a desired behavior is fast delivery of a minimum viable product (MVP) to catalyze mobile interactions, the organization must switch from reflexively treating \u201cdeviating from the plan\u201d as failure to celebrating \u201cfast failing or scaling\u201d as success.\nIf the desired behavior is \u201clearning by doing,\u201d then the organization must embrace the fact that harvesting data showing that \u201cwe were wrong and need to course correct\u201d isn\u2019t a proof of having made more mistakes: it\u2019s a sign of missing fewer.\nIf the desired behavior is minimizing risk, then the organization must embrace small changes that can be immediately rolled back as a stronger hand to play than approvals heaped upon gates piled upon process for large changes that can\u2019t.\nThese are a matter of both perspective and practice. To assess this idea, we ran a survey of IT decision makers across 800 large companies \u201cthe ability to meet or beat expectations\u201d with regards to app development cost, time to deliver, quality, quantity, and business impact of apps as a benchmark. \u00a0More than a third (36 percent) failed to exceed expectations on any of the five measures; 8 percent\u00a0exceeded expectations on all five.\nHow did they do it? For one, these \u201capp masters\u201d were far more likely than the rest to describe their IT department\u2019s approach as strongly \u201coutside in\u201d and \u201ccloud first.\u201d App masters also described their approach as \u201cagile\u201d and \u201cembracing experimentation.\u201d\nIn keeping with these attitudes, they were making greater use of cloud-based resources such as platform-as-a-service (PaaS) and infrastructure-as-a-service (IaaS) than weaker performers. And companies reporting the strongest capabilities using APIs and analytics were more than three times as likely to be app masters.\n\nWhile consumers have made the \u201cmobile mind shift,\u201d enterprises still have some learning to do. Among the hundreds of large enterprises we surveyed, only 1-in-12 are exceeding their goals for apps that appear to deliver on \u201cwhat I want, in my immediate context and moments of need.\u201d\nTo join them, the stimulus-response defaults for how your organization conceives, approves, funds and executes on digital projects and programs needs to change. As executives, strategists and implementers that ought to be an opportunity we\u2019re all salivating over.