We’ve both worked in technology for a long time, with internal and external clients, and a common challenge we’ve come across as data leaders is expectations. Not expectations like on-time go-live or on-budget delivery, but expectations of the impact data technology can have, and how quickly it can happen.

Here are two dynamics we’ve witnessed.

  1. Technology idealism
    New technologies like AI or machine learning clearly hold enormous promise. They are far from plug-and-play solutions, however. You have to identify areas of the business where you are seeking to augment and support human decision-making with machines. You have to create or integrate algorithms and test (some would say train) them thoroughly. You have to deploy them with operational data and ensure that they aren’t creating bias or change with data drift. You have to update them as conditions change. It’s part of the role of a data leader to inject some realism and strategy into the conversation, especially when a business executive or non-technologist leader has heard an idealized pitch of what’s possible. 
  2. FOMO
    Part of any technology consultant’s job is to assess trends and point to exciting new developments. That’s one way we manage the flood of innovation coming at us. Where it can go awry is when a business executive or board member expresses fear of missing out (aka FOMO) on the latest trend. Data leaders may feel pressured to jump onto new technology platforms before the decision is warranted. Here, too, it’s part of your job to get everyone in the organization walking before they break into a run. 

Five ways to bring realism to data analytics
Here are a few strategies that common-sense leadership can bring to temper organizational temptations.

  1. Pursue ambitious goals with realistic milestones. One thing we all learn pretty quickly working with technology is that you don’t run full tilt at a goal. You use spikes or labs to prove out ideas, explore edge cases, and seek to find points of failure – determining where value can be realized or quickly diminished.  Speaking to data culture, realism needs to come into play as it’s probably more useful to think of progress not as a high-speed sprint to your ultimate destination, but more of a long-distance run. Areas such as data literacy, data quality and data ethics need to be part of the long game plan, where the organization needs to place achievable marker points to guide change. 
  2. Expect change but be brave enough to challenge it. We’ve all been there. Your programs are finally running to plan, then your teams are bombarded with requests from key stakeholders who you thought were aligned with the strategic agenda. Your talent becomes exhausted and frustrated having to do context switching several times a day, and eventually they start looking for work elsewhere. If you as a data leader aren’t thinking about how constant priority changes have become part of the norm, you may be fostering an environment that causes your best talent to flee. If so, you need to get back to strategic lock-in and ensure that you’re not seeing diminishing returns from precious talent.
  3. Work through the potential implications of success. One of us recently did some work using machine learning and saw impressive results. The only problem is that we still had gaps in our operational processes, so we weren’t ready to take it forward to the organization as a whole. In another case, we generated a breakthrough result with data technology in a lot less time than we predicted. Here, too, we realized that if we tried to push it too far too quickly, it would cause complications and unintended consequences in other parts of the company. This is where a pre-mortem conversation is so helpful. If your project goes largely according to plan, will the organization be ready to adopt it? Better to think through the implications of success before you begin instead of after you barriers you could have predicted.
  4. Align data culture to business strategy. Fostering and nurturing data culture is a great way to promote realism about the power of data analytics. Make data culture part of the strategic themes of your organization. That way, your data efforts are progressive to realizing a strategy versus a sidebar that some might see as elective.
  5. Under-promise, overdeliver. Data leaders don’t need to pressure themselves into unrealistic expectations. Also, be confident in your convictions about where data analytics can help the business the most and deliver strategic impact for the organization. You can still under promise and over deliver, though. That’s a good strategy for nearly any scenario. 

The future really does arrive eventually
Five years ago it might have sounded inconceivable that you could do cloud-based analytics, for example, or create personalized life moment ‘nudges’ to better inform customers at the right time with data. Yet many organizations have persevered and arrived at these realities. It didn’t happen overnight, so there’s no need to promise that it will. Dialogue and education are the keys to building trust in your abilities. They also can create some pretty impressive results.