Demystifying Data Products: Essential Insights for Business and Technical Buyers

Overview

Data products are a hot topic. Everywhere you turn, someone is telling you that in order to get the most out of your data, you must manage it as a product that everyone in the business can use, rather than leaving it to become a chaotic mess of incomplete, inaccurate information that fails to deliver value. 

Sounds reasonable, right? But what, exactly, does it mean to manage data as a product?

Managing your data as a product requires your business to create a data product strategy implemented through the design and use of a data product. 

Still confused? Trust us, you’re not alone. 

While data products are getting a lot of buzz in the industry, there are many interpretations of what a data product is. And in our opinion, most of these interpretations are wrong.

Defining data products: let’s set the record straight

Data products make data tangible for everyone across the organization. They provide ready-to-consume sets of high-quality, reliable, and accessible data that people throughout the business can easily use to solve business challenges.

The best data products are organized based on business entities and governed by domain. They consist of comprehensive, clean, curated, and continuously-updated data sets that are aligned with key business entities, making them consumable by both humans and machines across the entire enterprise. 

Maximize business value with a data product strategy

Implementing a data product strategy through the design and use of data products is a surefire way for your organization to reap greater value from your data. Embracing a data product strategy delivers a myriad of positive impacts on the business, including: 

  • Better business intelligence: data products deliver the best version of your data for use in BI and analytics tools as well as in dashboards.
  • Exceptional customer experiences: data products help you understand your customers, inside and out, so you can power the experiences they’ve come to expect.
  • Optimized operations: validated, trustworthy data boosts operational efficiency and drives greater ROI so your business runs better.
  • Revenue gains: using data products, companies can spot new revenue opportunities by revealing previously-hidden relationships in the data.
  • Increased productivity: clean, trustworthy data improves the productivity of your data team, enabling business users to get answers to their questions faster. 
  • Agile decision-making: a single source of the truth eliminates data brawls because everyone is on the same page. 

What to know – and do – before investing in a data product strategy

Investing in a data product strategy is the best decision you could make for your business. And we’re not the only ones who believe this to be true. In their Hype Cycle for Data Management, Gartner talks about the transformational impact data products have on the way organizations manage data. 

What’s all the hype about?

As companies accelerate the movement of data to the cloud, new opportunities to innovate the way they manage data are emerging. Today, disruptive technologies such as data products enable organizations to uncover new insights from their data, empowering their business to work smarter.

Data products are a consumption-ready set of high-quality, trustworthy, and accessible data that people across an organization can use to solve business challenges. Using the perfect synergy of AI and human intelligence, data products reveal new insights that enable organizations to boost operational efficiency, power exceptional customer experiences, reveal untapped revenue opportunities, and safeguard their business from unforeseen risks. 

However, as Gartner cautions, failing to comprehend the business requirements and lacking a well-defined strategy to oversee the lifecycle of data products severely hampers their potential to deliver the transformative impact organizations need. That’s why it is critical for businesses to invest in a data product strategy and support it with an innovative data product platform. 

But before you jump into the deep end, there are four things you should know and do to ensure your organization is set up for success. 

  1. Review your data ecosystem

Before you invest in a data product strategy, it’s important to understand the state of your data union. Dirty data lurks in every corner of your data ecosystem, so it’s important to understand just how much clean-up you need to do. It’s also critical that you identify where data silos exist and make plans to break them down. Finally, take stock of how complete (or not!) your data is so you can determine the sources needed for data enrichment.

In addition, it’s critical that you understand your company’s stance on cloud providers. While it’s likely that your company has invested in a cloud data warehouse, is your company also fully-invested in their data stack? Hopefully not, because then, you have the flexibility to pursue a best-of-breed approach that enables you to select the right platform to support your organization’s needs. 

  1. Identify champions and secure buy-in

Everybody in your organization wants better data. However, not everybody is willing to invest the time and energy required to improve data quality. That’s why it’s important to find your champions early and leverage their support to secure buy-in and investment for your data product initiative. 

Look for business leaders who are passionate about making data-driven decisions. Show them how data products will help them uncover new insights from the data so they can work smarter. And engage them in the process of making data better. Once they see the value data products provide, they’ll become data product evangelists and will support your proposal to implement a data product strategy and invest in a data product platform.

  1. Evaluate your organization’s skills - and skills gaps

It’s likely that you’ve been investing in - or are starting to invest in - your data organization, otherwise you wouldn’t be reading this guide. And that’s a good place to start. However, there are new roles emerging that play a critical role in the successful implementation of data products. 

Before investing in a data product strategy, assess the skills within your organization. If you don’t already have a Chief Data Officer (CDO) in place, we suggest hiring one to lead the data product strategy, champion its adoption, and drive its success. 

We also recommend establishing a DataOps organization and hiring data engineers. Skilled data engineers are in high-demand, as they are the ones responsible for building the systems and infrastructure that collect, validate, and prepare the data for use by data analysts and data scientists. They also build and maintain data pipelines, and ensure compliance with their organization’s governance and security policies.

Finally, consider hiring data product owners, also referred to as data product managers, to be responsible for the development and success of your data products. This emerging role is proving to be a critical one. Data product owners act as translators between data engineers, data analysts, and end consumers. They collect the needs of the business, assign them a priority, and translate needs into requirements that the data engineers can develop.

  1. Review and understand the data product landscape

As you embark on the development and implementation of a data product strategy, it’s likely that you will want to evaluate data product platforms to complement the solutions in your existing data stack. However, just as there are many interpretations of what a data product is, there are at least as many vendors touting their “data product platforms.” 

It’s your job, as a business or technical buyer, to educate yourself not just on the definition of a data product, but also on the capabilities your data product platform needs in order to make your strategy successful. 

If this all sounds overwhelming, don’t panic. We’re here to help. We’ve compiled a list of questions you should ask data product platform providers when evaluating their solutions. 

This checklist isn’t like the general, watered-down list of questions you often find in buyer’s guides. We know that business and technical buyers need to understand different things to make an informed decision. So we’ve developed a list of questions every buyer cares about, as well as ones specific to technical buyers. Follow the path below to find the questions tailored to what you care about most. And if you’re curious, come back and explore the others, too!

  • I want to start at the top. Guide me to the general questions every buyer cares about.
  • I’m interested in the tech. Take me to the technical buyer questions.

[General Question] Questions Every Buyer Should Ask

As data products grow in popularity, so, too, do the number of platforms that claim to provide data product capabilities. Business and technical buyers have a lot of questions. And while some are specific to their lines of work, many cross the minds of both. Below is a list of core questions every evaluation team should ask when evaluating data product platforms. 

Hint: the best data platform providers will answer these questions from the point of view of both business and technical buyers!

Strategic Vision & Product Roadmap: What’s the plan - and how will you get there?

Data products are the new kid on the block. And while established data vendors offer innovative data product platforms, many new vendors are hyping their capabilities, too! 

The fact is, many of these vendors are unproven. That’s why your evaluation must include a frank discussion about strategic vision and product roadmap. Vendors in the space should have a clear view into not just where their platform is headed – but how they’re going to mature and deliver the capabilities they envision. They should leave you feeling confident that they’re investing in the right resources, right partnerships, and right technologies, not just jumping on the latest technology bandwagon. 

When evaluating data product platform vendors, ask them:

  • What is the strategic vision for your platform?
  • How do you see your platform evolving over the next 3-5 years?
  • What capabilities are on your product roadmap in both the short- and long-term?
  • How often do you deliver new releases on your platform?
  • How frequently do you update/change your product roadmap?
  • How do you gather customer feedback and incorporate it into your development process? 
  • What strategic partnerships do you have in place today?
  • Which data providers do you partner with for enrichment?
  • Are there any new partnerships or collaborations coming in the future that will enhance the platform’s capabilities?
  • How do you stay ahead of new and emerging industry needs?
  • How will you scale the platform as needs change and data continues to grow?
  • How do you stay ahead of new and changing security and privacy regulations?

Tip: Because data products are an emerging technology, it’s ok if a vendor doesn’t have all the answers (yet). The questions above, however, are table stakes. If a vendor waivers on their responses, that’s a red flag that they may not be the right partner for you.

Platform Adoption & ROI: What resources do I need and which use cases can I deploy?

Investing in a data product platform is a big decision and it’s imperative that you understand what you’re getting – and what you’re not.  Understanding what use cases the vendor supports today – and what use cases they plan to support the future is a good first step. Evaluate if these use cases align with your areas of greatest need. And perhaps most important, ask for examples of successful customer deployments and the ways in which they measure success. Bonus points if those customer success stories are in your industry.

You should also dig deep into the details so you understand what skill sets you need to deploy, manage, and use the solution. You may discover that you need to hire new resources or train existing staff to support the solution.

During your evaluation, ask the vendors:

  • What use cases does your platform support today? What use cases are you planning to support in the future?
  • How many customers are actively using your platform today?
  • How does your platform address our specific business needs and goals?
  • Can you provide examples of how your platform has helped businesses in my industry?
  • What ROI have your customers realized as a result of using your platform?
  • How do your customers measure success with your platform?
  • What skill sets should our organization have in place to support your platform?
  • What capabilities make your platform user-friendly for non-technical users?

Tip: Don’t be short-sighted. It’s important to select a partner that not only meets your needs today, but also has concrete plans to evolve their platform and deliver continuous innovation into the future.   

Onboarding & Support: How do I get started – and what will you do to support me?

It’s easy to get excited about the vision and strategy for a data product platform. But it’s important to understand the nuts and bolts of on-boarding and support, too. The best vendors prioritize customer support and invest in resources that help you succeed.

Make sure you understand how your selected vendor will support you before, during, and after your implementation. Ask them:

  • How do I get started with your platform?
  • Can you guide me through the on-boarding process, step-by-step?
  • What resources do you provide to help me and my team understand how to use your platform effectively?
  • What level(s) of customer support do you offer?
  • What resources or documentation are available for both business and technical users?
  • Are there training sessions or webinars available to new users?
  • Who do I contact if I have an issue? 
  • What is your typical response time?

Tip: Ask to see samples of on-boarding resources and documentation so you can assess if the level of detail the vendor provides aligns with what your organization needs. 

Strategic Insights & Analytics: How do I assess the quality of my data?

Improving data quality is a goal for every data product platform deployment. So how do you assess if the quality of your data is improving? The best data product platforms provide insights that help you gain greater clarity into the state of your data and how it evolves over time.

During the evaluation, it’s important to understand what insights and analytics the platform provides and how users, both business and technical, can access them. Now is also the time to ask how users collaborate and provide feedback to make the data even better. Ask the vendors:

  • What is your platform’s strategy for providing strategic insights and analytics?
  • How do we access insights and analytics within the platform?
  • What types of insights and analytics can we expect to get from the platform?
  • What integration capabilities exist to deliver comprehensive insights?
  • What capabilities exist to help business and technical users collaborate within your platform?
  • How do business users provide feedback on data quality?

Tip: Make sure you are clear on the level of technical know-how business and technical users need in order to collaborate, access insights, and provide feedback. The best platforms make it easy for everyone across the business to consume data and provide feedback on it via a closed-loop process. 

[Technology Question] Questions That Matter to the Tech Buyers: Let’s Geek Out!

Below is a series of questions and considerations that enable you to look under the hood of data product platforms so you can understand the inner workings of the technology and the skills your organization needs to run them successfully.

But before we get to the good stuff, there are a few more questions you should pose to your team to truly understand what you’re looking for in a data product platform. The big, age-old question that will drive your next steps is will you build – or will you buy?

Answering this question requires thoughtful reflection. While building may sound like a feasible response, in reality, it’s likely not the most efficient – or effective – use of time for you or your team. To answer this question confidently, you must first determine the following:

  • Is your company going all-in on a cloud provider’s data stack? 

Many cloud providers are slowly building out complete data engineering platforms and tying them into their marketplaces via pre-built templates. If you have the right talent in-house, it’s possible to use these offerings to do some amazing things. The key, however, is knowing whether or not you have the right talent. It’s unlikely that your team will have the skillset to build a great data product platform that works on multiple cloud platforms without taking a best-of-breed approach and building on top of other PaaS offerings.  

  • Do you know which cloud your company is going to use in the future? 

If you need to preserve flexibility, you’ll need to pay a premium. Going with a best-of-breed approach, one where vendors offer platforms across multiple cloud providers, will cost a bit more. However, if having the flexibility to move your workloads to another cloud matters to you (and for many companies, it does), then it’s likely you’ll want to follow this path. Because investing a bit more today will give you the flexibility you need tomorrow and beyond.

  • What experience does your team have with different vendors and toolsets? 

Start by understanding if you have cloud engineers ready to own your Amazon Web Services (AWS) or Google Cloud Partners (GCP) infrastructure using the cloud provider’s solution. If your team doesn’t have the skillset to do this, then going all-in on a cloud provider’s solution may be a bridge too far. While the cloud providers will tout their service as “easy to use,” as soon as you attempt to deploy it, you’ll realize that your team needs to know how your virtual private cloud (VPC) environment works, what level of permissions you’ll need, and the type of hardware to support it. 

To put a finer point on it, below is the list of skills that are must-haves if you want to build your own platform using a cloud provider’s solution. If your team checks all the boxes, then building your own solution is an option. If you don’t, then we suggest it’s time to consider buying a best-of-breed platform that’s easier to use. 

  • Cloud engineer: someone who knows Terraform or an equivalent like CloudFormation, has the skills to set up a full cloud environment and manage your data warehouse or lakehouse, and has proven experience with tools like Snowflake, BigQuery, or Redshift.
  • Data engineer: this individual will build and maintain your data pipeline. At the very least, they must know SQL and have some level of Python experience for the orchestration.
  • Analytics engineer: someone who has experience with BI tools such as Qlik or Tableau and can transform the cleaned-up data into a format that stakeholders can use.

  • Do you need a platform that supports multiple data products – or do you need to solve a single, specific problem? 

Now is the time to think about the big picture. While many businesses start by solving a single, specific problem, the reality is that in the future, they’ll likely have more issues they need to tackle. Solving a single problem is possible with a jack-of-all-trades data engineering team that is using best of breed tools that integrate well together. However, if you anticipate needing to solve more problems in the future, it’s worth investing in a platform that provides more leverage and reusability. 

If, after careful reflection, you’ve decided to build your own solution, then your journey ends here. But if you’re smart (and we know you are!), you’ve likely realized that buying a best-of-breed data product platform is the better, more efficient path forward. 

So let’s get to the part you’ve been waiting for. Below is a list of questions to guide your conversations with data platform providers so you can make the best decision for your business. 

The Fundamentals: Do you cover the basics?

While these questions are not specific to data product platforms, they are important questions to ask when evaluating any technology platform.

  • Describe your security posture. 
  • How long has your company been in business?
  • What compliance certifications do you have in place? 
  • Will you support the company’s IDP - or will we need to pay a SSO tax?

Buyer beware! When evaluating vendors, it’s important to understand if the company you’re considering is well-established (or not!). Data products are a relatively new concept, and some vendors are very much in start-up mode, hitching their wagons to the buzz around AI and large language models. So do your homework. Check their security certifications. And make sure you feel confident that they’ll be around for the long term.

Data Locality: How will you store, process, and protect my data in multiple regions?

Many companies today operate in multiple regions. And each region has its own set of rules, regulations, and compliance requirements. Established data product platform vendors should know what it takes to store, process, and protect your data in multiple regions, but asking these questions to newer vendors in the space may reveal vulnerabilities. 

Ask the vendors:

  • Where are you storing and processing data? 
  • Do you have data centers located in the EU? UK? US? 
  • Can you support multiple regions? If yes, which ones?

Tip: Cover your bases! Make sure you ask about every region you operate in today – and any regions where you might expand operations. That way, you can ensure that the vendor you select can support you now and in the future. 

Modern Data Stack: What technologies are part of your data product platform?

Understanding the underlying technologies within a data product platform is a critical part of the decision-making process. Every vendor approaches technology differently and has different requirements or limitations to consider. Not only will their approach determine the capabilities within the platform, but it may influence the resources you need to support it as well.

As you evaluate data product platforms, make sure you understand the details around their modern data stack and the skillsets you need to support it. Ask the vendors:

  • Do you require your vendors to be on a specific cloud? Are there any clouds your vendors can not be on?
  • Does your organization have a standard data toolset? What examples can you provide to help my organization get started with them? Examples include:
  • Visualization: Qlik, Tableau, Looker
  • Data warehouse: Snowflake, BigQuery, Redshift
  • Describe your modern data stack. Are you using technologies such as Spark or Beam to power your processing? Are you using SQL all the way down?
  • Are you cloud-native?
  • Is your billing consumption-based? 
  • What happens if we need to increase the frequency that we run our pipeline - from weekly to daily, for example? 
  • What if our stream throughput goes from 100 requests/minute to 10 requests/second?

Buyer beware! Many data product platforms take into account a number of assumptions when setting your cost. Changes to these assumptions will likely change the cost, sometimes dramatically. When talking with vendors about total investment, think beyond what you need today and understand how your pricing could change as your adoption grows. Understand all the supporting technologies you need to run the data product platform, and make sure your investment summary captures every line item.   

Machine Learning: Where do humans fit in?

Human intelligence plays a crucial role in reviewing data within ML models to create high-quality training data. Humans effectively guide ML in curating data for optimal output. The partnership between ML and human expertise ensures that training data fed into AI models is of the utmost quality, resulting in more robust and accurate AI/ML systems. Vendors who embrace AI/ML with human feedback in their data product solutions help maintain high-quality data, substantially boosting AI/ML systems' performance.

When evaluating vendors, make sure you fully understand how machines and humans work together to deliver the best data possible. Ask the vendors:

  • Is your platform machine learning native?
  • How does your platform incorporate human feedback?
  • Describe the process for humans to provide feedback. 
  • What capabilities exist within your platform to make providing feedback simple and straightforward?

Tip: It’s important to understand not just if data consumers can provide feedback, but also how they do it within the platform. Ask the vendor to walk you through the process, step-by-step, so you can feel confident that it’s simple and intuitive for end users. 

Generative AI (GenAI): What practical applications of LLMs are in your product today?

GenAI is taking the world by storm. And many data product vendors are touting their use of large language models (LLMs) within their platform. LLMs have the power to provide transformative capabilities that accelerate an organization’s ability to deliver clean, trustworthy data. But during your evaluation, it’s critical that you dig deep to understand the role LLMs play in their current and future product strategy.

Ask the vendors:  

  • How do you use large language models (LLMs) today?
  • What is your vision for how your usage of LLMs will evolve into the future?
  • How do you apply the use of advanced AI to compare and score diverse data sets?
  • What is the role of advanced AI in the referential matching of internal and external data?
  • Do you use advanced AI for probabilistic and deterministic matching?
  • Do you use LLMs for semantic comparisons?

Buyer beware! LLMs are all the rage, but many vendors offer little more than hype. When asking about a vendor’s use of LLMs, don’t just ask them about their vision. Look for tangible examples they can demonstrate today such as using a LLM to categorize resources to a standard taxonomy. If all they offer is a chat interface for asking questions about the data, then they likely just bolted on the feature to say that they are leveraging LLMs – and this is likely little more than marketing hype.

Data Enrichment: How does it work?

Most data product platform vendors support data enrichment in some way, shape, or form. Yet few of them do it well. Because the goal of a data product is to deliver the best version of your data, it’s important that your data product platform enriches the data with the best possible source. Many times, that source is outside of your organization. 

Understanding the ins and outs of how a data product platform vendor enriches data is critical. Without the right enrichment capabilities, your data will suffer and remain incomplete, incorrect, and outdated. Ask the vendors you’re evaluating to answer the following:

  • Do you support enrichments or the ability to bring in external data to augment our own data?
  • What partnerships do you have with external data providers such as Dun & Bradstreet?
  • How do you validate phone numbers and addresses?
  • How do you add an extra data source to the data product?

Tip: When it comes to data enrichment, size matters. Look for vendors with a rich database of companies and a robust set of partnerships so that you have access to the best sources of enrichment for your data. 

Platform Updates: What does it take to maintain the platform?

While not as exciting as learning about the bells and whistles, understanding how a vendor updates their platform is critical. Now is the time to ask granular questions about the how – and how often – you need to apply updates. It’s also important to know what happens if your environment changes (which inevitably it will).

Ask the vendors:

  • How do you update the data pipeline?
  • How often is source data updated? How often does the outputted dataset need to be refreshed? And how does this happen – batch vs. streaming?
  • How do you update the infrastructure and tools that power your data product?
  • How will the system scale to accommodate our projected data growth?
  • How do you handle schema changes over time? What happens if you need to add an extra column to your source table?
  • Does your solution address a single entity or problem – or is a platform for data products?
  • How does your platform handle cross entity/data product linkage?

Tip: Describe a scenario to the vendor and ask them how they address it. By giving them a real-world situation you’ve encountered, you’ll have a better understanding of the steps you’ll need to take to apply the updates using their platform.

Tamr: the leading data product platform

Tamr checks all the boxes when it comes to delivering the platform businesses need to deliver the clean, accurate, continuously-updated data that drives business outcomes. Our innovative data product platform is the first of its kind to unite AI with human intelligence to improve data quality and enrich data with first- and third-party data so businesses can revolutionize customer experiences, drive greater ROI, boost operational efficiency, and avoid risks. 

Using Tamr’s cloud-native and SaaS solutions, industry leaders uncover the insights they need to stay ahead of the competition in a rapidly-changing business environment. Don’t take our word for it, though. See how a global financial services firm is using Tamr data products to boost the bottom line by identifying and serving high-value customers better.

Seeing is believing: Tamr in action at Western Union

Western Union is transforming from a retail-first organization to a digital-first one. Their goal? To consolidate records from online and retail channels to form a 360-degree customer view. But when 200 million people used Western Union’s services to send and receive money in the past two years alone, this straightforward task immediately became very complex. 

Historically, Western Union relied on traditional, rules-based master data management to clean their data. However, this approach quickly fell apart when the company tried to connect large amounts of highly-variable data at scale. While these approaches helped them to clean 80%-90% of their data, it wasn’t enough to impact the bottom line. 

“You make a real difference in that 10% to 20% zone. That’s when you can reduce fraud or increase revenue. Our product has remained the same so to have a business result, we have to look at how much money we can make from those customers or how much risk we can reduce,” said Harveer Singh, Western Union’s Chief Data Architect.

Western Union knew they needed a better approach. Using Tamr B2C Customers data product, Western Union deduplicated and enriched 375 million customer records in a matter of months (not years), providing agents with a holistic, 360° view of the customer which allowed them to identify top customers, tailor experiences, and reduce marketing spend.

Read the full case study to learn more about Western Union’s journey to deliver better customer experiences using holistic, trustworthy data.

The data supply chain: a best of breed approach

With so many vendors talking about data products, it’s important to remember that, at their core, data products are reusable data assets that form the foundation for analytical and operational use cases. And to be successful, it’s critical that your organization has a high level of trust in these foundational elements. 

That’s why you should be wary of vendors who tout their data product’s ability to provide “everything you need” to address a specific use case. When considering the data supply chain, multiple tools have a role to play, from sourcing and transforming data to delivering it across the organization for consumption. 

Data transformation and accessibility are often most closely associated with data productsData products play a critical role in data transformation and delivery, but other, more technical solutions including metadata management tools, cloud data warehouses and data lakes, and data catalogs are important components of the data supply chain, too. Some require technical skills such as pythonETL or SQL, while others, like feedback & curation interfaces for data products, are more user-friendly, making it easy for non-technical users to engage with them. That’s why you should consider a best of breed approach. Together, these tools help data flow seamlessly across the data supply chain, empowering organizations to extract meaningful insights and drive innovation.

Choosing the right data product platform for your organization is a big undertaking. But when you understand the questions to ask and are aware of the potential “gotchas,” the process becomes a whole lot easier. 

Additional resources

Knowledge is power. When you’re evaluating data product platform providers, it’s important to be well-informed. So we’ve compiled a list of helpful resources to help you go deeper into some of the topics included through this buyer’s guide. 

🎧[Podcasts]

Finding the business value for moving data to the cloud, Evren Eryurek, Director of Product Management,  Google

Collaboration in data, Benn Stancil, co-founder and CTO at Mode Analytics 

Mastering your own destiny, Andy Palmer and Mike Stonebraker, co-founders of Tamr

Designing a product for making better data products, Anthony Deighton, General Manager of Data Products, Tamr