Choosing an embedded analytical and reporting tool isn’t like choosing furniture. It’s more like choosing a friend who’ll join you in a journey over the course of the next few years. A journey that, in a fast changing world, will be filled with unknowns and challenges.
What is Embedded Analytics? Gartner defines embedded analytics as “The use of reporting and analytic capabilities in transactional business applications. These capabilities can reside outside the application, reusing the analytic infrastructure built by many enterprises, but must be easily accessible from inside the application, without forcing users to switch between systems.”
We can go a step further and include the ability to embed into any cloud solution and/or product, as well as keeping a consistent user experience when switching between systems. Embedding is not only about sharing the same authentication (single sign-on), but also about having a consistent and coherent look and feel for all the systems involved in the solution.
In this article, we’ll share our experience as a provider that helps customers embed icCube into their solutions.
Embedding a Tool is a Journey
Some companies are able to complete a project with the same set of requirements for years. But products and solutions evolve quicker and quicker, and your embedded analytics module will have to follow this evolution. Over the past four years, only one icCube customer has run the same version they initially launched with in a 24/7 production environment without installing any updates. A robust integration like this was only made possible because of the great teams, months of testing and having a complete list of detailed requirements.
You might have this as well, but this level of detail is often an exception. Most likely, you’ll find yourself working closely with your vendor, not only during the integration phase, but for the next few years. In this journey, like when trekking in the middle of a mountain when bad weather hits, you’ll appreciate being with somebody you can count on. Therefore, choosing a high quality partner is an important factor to ensure you’ll have a smooth journey.
You can plan a sunny and relaxed journey, but you’re really safe when you know you can rely on your partner to handle the bad times. These times happen even to companies that deliver high quality products, so choose a partner you can make your journey with; a partner that shares your values and is flexible enough to align with your goals.
Extra Mile 1: Development Tools
Development tools are a set of tools that developers use when coding. They range from a smart source code editor (code completion, error detections, etc), a compiler and, eventually, a debugger. These aren’t your average tools to play with; they’re used to shorten development time by an order of magnitude, and improve quality.
No professional development projects can be done without a set of development tools. There’s no reason why business intelligence should be any different. Don’t be tricked by sales messages. Unless you’re doing very basic analytics, you’ll need development tools. That’s why icCube’s R&D team cooked a unique MDX debugger, so business analysts can see how queries are solved by the engine step-by-step. This speeds up time to market, and enables you to deliver a better end result. So don’t forget to ask what development tools can be used with your new embedded analytics vendor.
Extra Mile 2: A Bigger Size than Initially Planned
You have the list of required features covered by the embedded analytics tool. You also know how the vendor reacts when you request a small (or not so small) change and it’s time for delivery. But can you plan for all your future needs?
An embedded analytical tool is something that’s going to stay for years, and should grow with your solution/product. Some projects have a fixed list of required features, we have a few like this in icCube, but most you probably won’t be able to plan for years in advance. If your users have access to the analytical layer and can create reports, they’ll ultimately gain experience with the tool, and possibly request more advanced features on both the analytical and reporting sides.
Therefore, we advise choosing a tool that can grow with your business. Check how easy it is to extend features in the reporting tool and check if you can create your own widgets. Ask yourself: is the query language powerful enough? Can I combine it with other languages like R, Java or C#? Eventually, you’ll need an embedded solution with a lot more horsepower and a lot more widget flexibility than you initially planned to cover your future growth.
Extra Mile 3: Predictive Analysis
Predictive analytics is the current trend in the Business Intelligence world, but is it something you should use? Indeed, any business can gain from better insights and forecasts. But you’re probably not Amazon, Netflix, Zara or a similar company that has millions of customers with a team of expert data scientists. Predictive analytics is great for customer behavior, but you’ll need a lot of customers to get meaningful statistics.
Doing statistics on a few events is mostly a waste of time. Zara has thousands of shops with millions of transactions, and it’s the same for other big names that extensively use predictive analytics. Those companies can conduct experiments to see how customers behave when the company changes how they sell (e.g. position of the products, colors, website, etc.) and they can get meaningful results in a few days. On top of all that, predictive analysis is a complex field. You’ll be better off if you hire a data scientist service or a specialized company.
Our advice is to start your analytics project by getting a better insight of your business, and then go for predictive analysis by looking for a field expert on both your business and data science. Indeed, you need to ensure that your embedded analytical tool supports advanced mathematical languages like R, and can handle advanced predictive analysis.