What are some effective strategies for soliciting and incorporating feedback from data users and clients?

To identify feedback goals and an audience for data management: 1-Align Objectives 2-Stakeholder Mapping 3-Diverse Audience 4-Priority Audience 5-Segmentation 6-Feedback Channels 7-Timing 8-Clear Objectives 9-Feedback Type 10-Incentives 11-Feedback Analysis These steps ensure that feedback collection aligns with data management goals and results in meaningful insights for improvement.

I go by 2 guiding principles when developing analytical solutions for users. 1. The end users typically don't know what they want until they see what they don't want. Most don't know what's available so they can't be specific about what the end result should look like. Use prototypes, wire frames, hand-drawn diagrams to elicit feedback. 2. Once you answer the first questions, there'll be many more. End users can't specify what all the future questions will be until they see the answer to the first question. This is the 'next question dilemma'. Consider "disposable" analytics - get used to throwing some of the analyses away once they've served their purpose.

I have found it much harder to seek feedback on data products, than on software. Feedback tends to be anecdotal and focused on instances where data could not be found, or had quality issues. For effective feedback, it is important to map the use cases. The same data service may be used for dashboards, AI/ML models etc. Gather feedback as close to the point of use as possible - e.g. at the end of Exec review meetings. I leverage the term "use-through" rate to measure data product effectiveness (as a rough equivalent of "click-through" rates in eCommerce). If the data product is found to deliver useful and accurate data with the right level of detail, it is most likely to be embedded and used more in analytical applications and AI/ML models.

In my experience, it is also critical to ensure feedback from your legal team. There are over 161 countries with data protection laws. Also over 127 specific AI use case laws are already on the books (eg employment AI in NYC). Further comprehensive AI laws are presently proposed in 37 countries and six continents. You will want to speak to experienced privacy, AI and cyber lawyers to ensure that your data practices align in this rapidly evolving legal landscape.

With the experience on data management and understanding its importance. Feedback comes way later however I would prefer to understand the data requirements first and foremost. Be precise about what is required and when and its storage, retention, retrieval and maintenance. Based on these first factors which does sounds extensive but that really helps with narrowing down the feedback. The feedback that can be handled and worked out rather than words posted and forgotten. Cheers, Ni

If you've built a data governance organization with representation from all functions in the company, then you've already identified the people you need to get feedback from. If you need to dig deeper in certain areas, you can have user groups of people with shared interests. You can also have a ticketing system, similar to a help desk ticketing system that most people are familiar with.

In addition, leverage technology platforms to automate the collection and analysis of feedback, making it easier to implement actionable insights in a timely manner. Tools like real-time dashboards can also be employed to monitor key performance indicators, ensuring that you're consistently aligned with your data management objectives. By systematically embedding the voice of the user in the heart of your data product development cycle, you not only fulfill current user needs but also establish a robust, self-sustaining ecosystem that continually refines and elevates your data products.

Time and again, I have found that the most useful feedback is from existing paying/regular/loyal customers. These loyal customer are the foundation and power-house for word-of-mouth marketing and so I have come to learn that how they describe the software to their peers is of utmost importance to your organization. Net promoter scores are quite good "Out of 10, how would likely would you recommend our service? And why?" but lately I have found the more active version even more effective: "Out of 10, how disappointed would you be, if you could no longer use our service? And why?"

Also, be mindful of who is responsible for the channel of feedback. You need to have someone/mechanisms that ensure feedback is received and processed as received. In the event that you have someone who is always eager to tell the good story, all that is perceived as negative will be filtered out.

Think of feedback collection as crafting a user experience. One size doesn’t fit all! The mode of feedback collection should reflect the comfort and preference of your audience. Whether it's a quick chat or a detailed survey, it’s crucial to provide an environment where data users feel their feedback is welcomed and valued. Remember, the method you choose is the first impression of how seriously you take their feedback.

Contradictory feedback can lead to extremely productive discussions. By openly discussing what prompted the feedback from each party, you may be able to uncover a common underlying issue and collaboratively identify a previously unknown solution that benefits everyone moving forward.

The devil is in the details - sometimes the actual challenge lies in separating your emotional attachment from the feedback. To prioritize effectively, it's essential to categorize the feedback based on criteria such as relevance, urgency, feasibility, and impact (nothing more, nothing less). Your goal is to spotlight feedback that tackles the most pivotal issues, delivers substantial value, aligns seamlessly with your data management objectives, and is feasible for implementation.

In my experience it is also important to gather feedback from stakeholders such as the legal department and compliance who can help future-proof your product development by making sure it is compliant with relevant upcoming legal trends. This is especially important for AI.

If your data is part of and serves a large operational model, with a wide user base, you can use the incident reports related to the operation, measure how much of the incidents originate from the data (integrity and availability, right?) and figure out from the discrepancies your best route to improve the quality of the data.

Data Product Feedback should be categorized and addressed according to criticality. For example, this is one order in which I have seen feedback categorized and addressed: 1. Accuracy of data product outputs 2. Accuracy of metadata on the data product in the data catalog 3. Completeness - are all relevant information elements available? 4. Timeliness of updates - is the output up-to-date? 5. Level of detail vs. summarization 6. Performance of the API or data product (speed) 7. Findability of the data product (e.g. is the API or Data product easily discoverable in the enterprise data catalog?)

It is also important to proactively seek further feedback after changes have been made. So the company is in the posture of affirmatively seeking feedback and not reactively awaiting follow-up.

This step is very important as data product changes are harder to discern for end users. The outputs of the products are consumed by other software and AI based applications. The enterprise data council and functional area data councils are a very powerful body that can communicate changes to business users in the respective areas and seek further feedback. Data API documentation and release notes are very important ways to communicate to technical audiences.

Analyzing feedback, developing an action plan, and communicating back to stakeholders should be done quickly. A quick turnaround with clear communication demonstrates that you value their feedback and take it seriously. This helps build trust and credibility with your audience and makes it more likely that they will freely share future feedback.

I strongly that feedback in the foundation for any changes, considering feedback is the essential steps to improving a ourselves, a system, an application, a process, procedures and our outlooks on the global environment. Once acknowledged, an action plan should be drafted as well communicating back an acknowledgement that the feedback was taken into consideration. Base on the nature of the feedback a priority is set for any type of mitigation. If the feedback was the building is on fire, the alarms would sound and the building evacuated, Every feedback has a nature of urgency. If feedback is not taken into consideration, sometimes the worst things can happen, preventative maintenance helps only if reports are consider & action taken.

Communication is a two-way street. Once feedback is received, the next step is to make it actionable. By sharing action plans with clients and users, businesses show transparency and trustworthiness. It's not just about implementing changes, but about building a collaborative relationship, where everyone feels they're part of the solution.

I have found continuous testing of data products to be very useful to understand improvements made base don feedback. Testing should cover usability, accuracy, timeliness, performance and "use-though" rates. I leverage the term "use-through" rate as a sort of a rough equivalent of "click-through" rates used in eCommerce. If the data product is found to deliver useful and accurate data with the right level of detail, it will most likely be embedded and used more in analytical applications and AI/ML models. The more effectively feedback is analyzed and implemented, the more widely and frequently will the data product be used.

Data standards, policies, and practices must adapt over time to remain effective. As data needs and industry trends evolve, having a deliberate feedback and improvement loop helps ensure that your company's data universe remains current and relevant.

When I find a significant issue, I exercise the feedback process and request the issue be addressed or escalated to the responsible team, In some cases escalated to management if appropriate. I subsequently request a follow-up if no actions are taken. Having feed back from all customers is important to improving all applications, systems, policies and procedures. Addressing and acknowledgement of mitigations is essential to customer satisfaction and company commitment to customers. Company footprints are small by comparison to the customers global foot prints. Take care of both!

Feedback is not a one-time event; it's a journey of continuous improvement. The value of feedback is maximized when businesses see it as an ongoing conversation. Every piece of feedback is a chance to refine, re-evaluate, and rejuvenate the data strategies in place. The more we listen, the better we serve.

One thing I have found helpful is that: Learning and improving from feedback is crucial because it helps us adapt, refine, and align our efforts with user needs and expectations, ultimately leading to better products and services

The optimal strategy for garnering feedback from both internal and external data users involves restructuring your data organization into a data product management-focused entity. By adopting this approach, data users become the core focus of all data product development efforts, eliminating the need for additional measures to solicit feedback.

Establish accessible and user-friendly channels, such as email, online forms, and suggestion boxes, to ensure clear communication with data users and clients for feedback purposes. Proactively seek feedback through surveys, interviews, focus groups, and user testing sessions at different stages of the data lifecycle to gather valuable insights and improve data products and services.

AI legal frameworks for trustworthy AI, contained in draft legislation in 6 continents will require companies to have high quality data for “high risk” AI (eg AI for financial, health, children, critical infrastructure and 50+ other areas). This means that if your company is licensing a large language model from a tech company that care should be taken for the data your company uses to train your proprietary application. Data will need to be managed and companies will want to have easy ways of identifying the data to used to train their own proprietary use of the customized application built on top of a large language AI model.

When you encounter a rock in the path, do like water flow around it. Go further and reach out to other products to determine if you encounter the same limitations, if not you can then choose your options. Change is the one constant in the universe. It's great when things change for the better and leverage feedback to improve things going forward.

Often, the feedback that isn’t straightforward is the goldmine of insights. For instance, a story about how a user navigated through a data tool or an offhand comment about data accessibility can provide profound understandings. Encourage anecdotal feedback, encourage stories, and you'll gain not just data points, but a narrative of the user's journey. In conclusion, feedback in data management isn't just about improving products and services; it's about building connections, understanding stories, and co-creating a data-rich future. Embrace feedback, and you embrace growth.