Embedding analytics — build vs. buy

An estimated 40% of companies already embed analytics in their customer-facing applications. That number is only increasing as customers demand data and companies see return on their analytics investment to the tune of stronger competitive differentiation and more satisfied users.

However, not every company has the technical expertise to build really great analytics into its products, services, and experiences. So how are so many companies able to embed analytics so successfully? A few managed to build the capability themselves, but many chose to buy and embed an analytics solution. This report will help you understand the pros and cons of the two approaches.

At the heart of this decision lies an assessment of your organization’s R&D or IT capabilities, then understanding the risks and benefits that arise when relying on internal development teams.

40%: Proportion of organizations embedding analytics

25%: Proportion of analytics capabilities that are embedded in business applications

Many companies today embed analytics into their applications so users can access and act on insights from their data. The prospective value is twofold: happier, more successful customers, and a strong competitive differentiation.

Learn all about the benefits and disadvantages of embedded analytics and buying versus building, and explore the pros and cons of each, then get ready to ask yourself the important questions listed in the last section as you grapple with the big decision: Will you buy or build your analytics solution?

The pros of building analytics (in-house)

Your customers want analytics. Companies that begin their analytics journey with an in-house solution usually start with user (or customer) requests. Most products and services today come with some way to show data to the user. As users become more data savvy, their appetite for more data and more insights grows. Staying competitive in the market today means you need to provide not merely good, but great analytics for your customers/end-users. For example, notifying users of outliers and sudden changes in performance and forecasts that indicate future problems or showing the correlated data that shows a drop in usage.

When it comes to building analytics into your product, engineering teams usually jump at the chance to try something new, and analytics visualizations present a meaty challenge for building and integrating something new inside a familiar infrastructure.

Soon, users are viewing eye-opening information that is embedded, customized, and integrated into the business apps, giving them an exceptional user experience. Be ready; they will be coming back for more!

User experience is critical: By building the analytics functionality in-house, you retain complete control of the look and feel — defining who sees what, when, and how.

Still gauging market needs: You may not yet know if your customers need sophisticated analytics or just basic reporting. Rather than invest right away, you can start building and see how it goes.

Engineering passion: Building analytics functionality can be exciting for your team. (We get it!) You may have reason to use the excitement as momentum rather than risk discouraging them.

Cost efficiency: Putting a couple developers on charting for a sprint or two may be comparatively inexpensive. Plus, you won’t have to build a business case or justify the decision.

The cons of building analytics (in-house)

Building analytics seems simple at first but is more complicated than many think. Building the V1/minimum-viable-product version of your analytics might not be a huge challenge, but the ongoing maintenance that will be required is an important aspect to consider when making the decision to build instead of to buy.

While building in-house gives you complete control over features and integration, it can quickly escalate into a long backlog of requested features, performance issues, and heavy demands on an entire development team to keep up with growing needs. Also consider how the actual analysis and insights are written. It is likely that developers will need to be roped in every time a product manager wants to roll out a new metric to measure, or an analyst wants to prototype small changes to reporting.

This requires ongoing investments of time and budget from R&D — a budget that is likely to get exorbitantly high as data, users, and overall analytics needs increase.

Basic visualizations don’t stand out: Your team may have built a prototype in a week, and promised Version 1 in a quarter. Consider: how long will it take to get to Version 2, or to capabilities considered table stakes for BI tools: cross-filtering, drilldowns, PDF and XLS downloads, pivot tables, notifications, alerting, anomaly detection, custom queries, or machine learning?

What does the next chart cost: In-house developed analytics are generally hand-coded, rather than relying on general purpose frameworks. How many sprints will it take to get the next set of insights into the product? What about changes and improvements to existing analytics — will you need engineers involved at all stages of development?

Scaling the platform won’t scale analytics: Running queries and charting results may work efficiently in a prototype. However, as your customers become reliant on your analytics, they’re going to create a whole new set of performance problems that your team will need to solve. The problems likely will be unrelated to the scaling challenges you’re already addressing in your own platform. Will you be able to invest in a dedicated infrastructure team to keep the analytics running?

R&D is an expensive resource. Analytics and BI projects are long term and require constant iterations, resulting in a serious cost as you divert expensive R&D resources to non-core activities like building and maintaining an in-house BI solution.

Another issue that arises from growing BI out of your technical department is that the knowledge of the actual BI needs stems from the business users rather than the technical departments. This leads to significant risk that the BI solution will be misaligned with actual requirements and need constant and time-consuming iterations.

Finally, as the amount of data and number of users grow and changes in your company arise, an in-house solution will need to develop quickly to meet new, unanticipated needs. Due to differences in priorities between R&D and business departments, users often experience excessive delays in getting the data or analysis in a time frame that is actionable and fail to deliver results.

Remember, an established analytics provider grows and innovates with the market, and an in-house BI tool will require time and R&D investments you can’t really hope to match.

Smart building/Energy management platform Vitality wanted to elevate its product to enterprise-grade with the help of embedded analytics. Customer demands had grown beyond what Vitality could build in-house and they needed to buy the analytics capabilities to stay relevant.

With Sisense, they had an efficient, cost-saving analytics solution that would perfectly match the look and feel of their software, which enabled it to go to market faster and quickly add value with pre-built features and open-source APIs that integrated seamlessly with the existing products.

“Our core strength is algorithms,” says Clayton Erekson, Vitality CEO. “We hadn’t put resources and development behind data visualizations. I’m guessing we would have spent millions of dollars trying to build it ourselves.”

Buying analytics rather than building its own solution not only reduced costs for Vitality, but also enabled it to test new features, and landed them enterprise clients.

The pros of buying analytics

Today, it’s much more common to embed an analytics and business intelligence solution by working with a BI partner and embedding its product through an original equipment manufacturer agreement. Embedding is proven to be the easiest and most effective way to offer business intelligence because it delivers a faster time to market by using an established BI solution and technology.

Many of the resource allocation and budget issues dissipate by embedding a BI solution, especially one with technology that can easily scale to your current and future data needs.

Get to market faster: Meet customer needs faster without having to build from scratch. Empower customers with self-service analytics and augmented insights.
Save on costs to build and maintain: Avoid spending on development resources required to build and maintain a full-blown analytics solution.
Deliver innovative data experiences rapidly: Deliver cutting-edge data and analytics capabilities including AI- and machine-learning-powered features like natural language querying without the need to build in-house expertise.
Stay focused on core problems: Engineering and R&D can focus on the main business problems they are trying to solve instead of trying to build analytics.

BI analyst and researcher Wayne Eckerson says the future of BI and its continued success in terms of user adoption and extensive deployments are dependent on embedded analytics.

“The best way to simplify and operationalize BI is to embed it directly into operational applications and processes that drive the business. This is the definition of embedded analytics, and it’s the next wave in BI.”
BI Analyst Wayne Eckerson

There are plenty of self-service analytics solutions in the marketplace. It’s useful to consider to what extent they can help make analytics a part of your daily work routine. Sisense stands out with its ability to infuse analytics into your workflows. Sisense ties into third-party systems (like Google Ads) to present users with actions to perform right next to the insights, and the ability to use data (like current ad spend, for example) to adjust settings in other systems.

The cons of buying analytics

In some cases, buying an analytics solution might be too big a leap for your company. Or, maybe your core business somehow includes data and analytics — then it might not be too much of a stretch for your R&D and IT teams to support building and not buying an analytics solution. Those two points could be reasons not to buy, and here are a few more points to think about.

Can you get enough money to support an analytics solution? An enterprise-grade solution can require a full budget approval process, maybe including RFIs and RFPs, and continued budget support throughout the years for new licenses and upgrades.

Getting and maintaining budget: An appropriate budget should be allocated for an analytics solution, including ongoing upgrades and new license expenses.
Securing buy-in: It’s important to get buy-in from developers and engineers, as they are the ones who will be implementing the solution, and embedding analytics needs the full support of management.
Digital transformation plan: Analytics are a catalyst in the growing cloud ecosystem, which requires a well-planned transformation strategy (Gartner report).
Legacy data systems: In order for modern analytics to be effective, data connectivity to legacy systems will need to be established; silos and older technologies may be a blocker.

Engineers are often the biggest advocates of building a solution, and if you are leaning toward buying, then you will have some work to do getting engineers, IT, managers, and other team members on board. All will be crucial when implementing the solution.

Making the build vs. buy decision — questions to ask

In the buy versus build debate, it’s critical to ask yourself the right questions and support your decision with the metrics to make your case.

Is analytics core to your product and strategy?

If analytics is your core business model, building could be an option, granting you full control over the development.

Can you build it well? (Do you have the expertise?)

If analytics is core to your business and your strategy, do you have the expertise to build and maintain an analytic application? Building analytics is more than creating charts over a database query. A sophisticated analytic application requires high performance, scalability, speed, security (including role- and user-based access) besides innovative data experiences. Ideally, you need resources with domain expertise in building and maintaining analytic applications for successful launch with a build-from-scratch strategy.

Do you have the resources and time?

If expertise is not an issue, then the next question is if you will have the development resources. This will require either reallocation from other teams or new hires. Building from scratch will require dedicated engineers and teams. Additionally, finding top-tier analytics developers and engineers will be challenging.

Secondly, what is the timeline for launch? Is it three months or one to two years? The shorter the launch timeline is, the more difficult it will be to build a solution completely from scratch in-house.

Can you effectively maintain and grow the solution over time?

It might be easy to put together a starter dashboard with a few charts that run queries against a database. But what happens when there is a new question that comes up from customers? Will you have the resources to go back and build those new requirements in? What happens when the data size and complexity grows? How will the team maintain performance with the complex queries? What happens when the number of customers and users grows and the concurrency load goes up? What happens when there are new requirements that crop up more often than ever? How do you support all these changes? It is clear from this list of questions that maintaining and growing can be a difficult task for internal teams.

Is it cost effective?

Perhaps the most important question. Given your investment, will building in-house make sense from a cost perspective?

Delivering a world-class data experience that incorporates rich, interactive interfaces with fast time to insights takes a team of developers. Basic math will quickly show that this cost can be substantially higher than buying. Two developers working full-time on the project is already $300,000. In addition, they might take a year or more to launch a full product. On the other hand, the buy option typically requires fewer resources and for shorter periods. A break-even analysis that takes into account the licensing costs, developer costs, and revenue accruals across a few years will quickly show that the return on investment with buying will quickly outweigh building.

If the answer to all of the above is a resounding “Yes!” then build.

If the answer to any of the questions is a “No,” then it is better to buy a solution.

Buy with the ability to adapt and extend

When deciding to buy, it is important to consider how much you will need to expand in the future and maybe avoid solutions that are closed boxes. Customer requirements are ever changing. In addition, the number of users, data, and use cases can grow without an upper limit.

The platform that you partner with should be able to meet these changing requirements through a flexible, customizable, and scalable platform. To help you decide, ask yourself the questions:

Do you need to provide a seamless customized experience?
Do you foresee unique needs with more customers and use cases?
Do you see automation and integration needs in the future?
Do you anticipate changes over time?

If you answered “No” to any of these questions, then you are probably safe with buying a closed solution that will suit your immediate needs with no expansion plans. If the answer to any of these questions is a “Yes,” then it is important to buy an open platform that can be extended, automated, and scaled easily.

Partner for success

Thousands of customers have already weighed in on the Sisense pros and cons and are partnering with us for embedded analytics in their products and services.

Together we:

Drive customer satisfaction and product stickiness: Embed intelligence exactly where your customers need it with the exact experience you want to provide.

Build future-proof products that are secure: Ensure maximum flexibility to support growth and change.

Scale without compromising performance or agility: Scale programmatically as your users and use cases grow.

Manage your data at scale: Deliver real-time insights with live data or cache data for rapid prototyping and minimized query costs.

Stay ahead of rapid innovation cycles: Get to market fast without a heavy engineering lift, enabling customer self-service.

Deliver deeper insights: Leverage code-first capabilities for custom and unique needs.

Stand out in highly competitive markets: Deliver innovative user experiences that go beyond standard analytics.

Customers partnering with us: DNV, GE, and Geneia

DNV, an assurance and risk management company, is generating new subscription revenue by embedding Sisense in an analytics platform for utility companies. This platform, Cascade Insight, is purpose-built to deliver actionable insights about utility asset performance. Its customers are able to get ahead of maintenance issues and avoid service outages in their grid, ultimately saving money and delivering better value and service to their customers.

GE, an American multinational conglomerate, uses Sisense across more than 150 hospitals in an application called GE Smart Scheduling. Teams are able to predict no-shows and cancellations for MRI and CT scans. With Smart Scheduling, one imaging center in Chattanooga was able to see an increase of $100,000 in revenue per year based on decreasing no-shows and cancellations, thereby providing better care to its patients and communities while optimizing machine usage.

Geneia, a healthcare analytics and services company, uses Sisense and AWS in its Theon Platform to help hospitals, doctors, and providers to drive more personalized, patient-centered care to over 7 million people. Embedding Sisense helped the company provide more stickiness in its product as well as improve the overall healthcare process. The CEO has stated that “Sisense has been a game changer for our organization.”

Ready for more? Use the following resources to help find the best solution for you:

Compare Leading Embedded Business Intelligence Software with G2 Crowd to ensure that the platform is customer vetted.

Compare Now

Pros & Cons of Building vs Buying Embedded Analytics