Transcription

4 Steps to SuccessfullyEvaluate BusinessAnalytics SoftwareIntroductionThe goal of business analytics and intelligence software is to helpbusinesses access, analyze and visualize data, and thencommunicate those insights in meaningful dashboards andmetrics. Unfortunately, the reality is that the majority of softwareoptions on the market today provide only a subset of thatfunctionality. And those that provide a more comprehensivesolution, tend to lack the features that make it user-friendly.With a crowded marketplace, businesses need to go through a complex evaluationprocess and make some fundamental technology decisions before selecting a vendor.Finding a business intelligence (BI) software that will scale with your organization’sneeds may seem like an impossible task. Here are the four questions you can askwhen beginning the BI evaluation process that will save you a lot of time and help setyou in the right direction.www.sisense.com

1. How Do I Select A Software Stack?To answer this question, first you need to better understand theoptions for extracting intelligence from your data. It is important todifferentiate between business analytics and dashboard reporting.Business intelligence (or business analytics) is a set of theories, methodologies,processes, architectures and technologies that transform raw data into meaningfuland useful information for business purposes.By contrast, dashboard reporting projects have a more limited scope and generallyaddress current requirements rather than future ones. Dashboards and reports tendto be static, created once and simply refreshed with updated data, as opposed tobusiness analytics, which allows an organization to create new reports anddashboards as required.Project-Specific vs. Solution-Oriented ApproachesOrganizations that choose a project-specific approach generally require a short-termsolution or have projects of limited scope. Reporting requirements are predictableand consistent, and can be handled with static reports based on single or simple datasources.Organizations that have a broader view of their requirements will use a longer-term,solution-oriented approach. These organizations understand that their need forreporting will get more complex as data accumulates and gets more complex, and asthe rate at which new reports are introduced will increase over time due tointegrations, customizations, time for training and increased adoption.Businesses that opt to use a solution-oriented approach need software that takes intoconsideration current and future reporting requirements. The most effective way tohandle this in terms of technical architecture and scalability is by implementing a datawarehouse and possibly data marts. (See below).www.sisense.com

Technical OverviewIf you take a forward-looking approach and see that your business analyticsrequirements will develop over time, implementing a data warehouse architecturewill allow your solution to scale with your needs. That being the case, your businessanalytics architecture will probably look similar to the diagram below. Data from alldata sources is pulled into a centralized database through a process calledExtract-Transform-Load. This databased is called a data warehouse (DW).DATA WAREHOUSE(SQL, Redshift,Snowflake, REDA data warehouse should be viewed as your company’s ‘single source of truth’ bycontaining a compilation of only clean and accurate data. Data warehouses can househistorical data as well as current data.Take note: Management of a data warehouse is oftenhandled by the IT department, limiting the access of allother business groups within the company. Instead,individual teams or business groups may access data inthe data warehouse via data marts.www.sisense.com

A data mart is a subset of the data warehouse that is oriented to a specific businessgroup or team. Whereas data warehouses have an enterprise-wide depth, theinformation in data marts usually is limited and only pertains to a single department.Both data warehouses and data marts refer to the back-end of the business analyticssoftware stack. This is where the data is stored, transformed and managed. Neitherthe data stored in a data warehouse nor that in a data mart is accessible by end users.In order to extract the data from the back-end, you will also need a front-endvisualization tool for data discovery and visual analysis.Different Types of Software StacksFront-end stackDATA WAREHOUSE(SQL, Redshift,Snowflake, etc.)DATADWETLRESULTSBack-end iness analytics software is offered in three different configurations.Back-End Software Stacks: Provide only back-end functionality such as datastorage, transformation and management (i.e., data warehouse and data martfunctionality, as well as ETL capabilities).Front-End Software Stacks: Provide only front-end (end-user facing) functionality,such as data visualization and visual analysis.Full Software Stacks: Deliver both back-end and front-end functionality.www.sisense.com

Full Stacks vs. Front-End StacksIf you have not gone through the process of collecting, centralizing and standardizingyour data, using a front-end stack instead of a full stack has several negativeimplications:Front-End StackFull StackDetailed DataDetailed DataDirty DataClean DataNon-CentralizedCentralizedData Mart Capacity or Excel/CSVData Warehouse CapacityThe advantages of using a full stack as opposed to a strictly front-end stack are:Enables centralized data storage and managementAllows for consolidation of multiple data sourcesScales to far larger data volumes and more concurrent usersMinimizes the number of data marts to develop and manageAvoids need to create and work with multiple (and potentially huge) CSV extractsProtects operational systems from damage (rogue queries)Allows access to data sources that cannot be queriedWhen Would You Need the Full Software Stack?In most cases, implementing only a back-end stack or a front-end stack would notsuffice to ensure a real, effective and scalable business analytics solution. In particular,you would need a full stack when:End users want to access centralized data and maintain a single version of the truthrather than build a solution around CSV/Excel extracts.ETL functionality is required, which often happens when multiple data sources areinvolved and when the data is especially dirty or especially large.The data sources cannot be directly queried, either because they are not supportedor because they are part of a critical operational system.www.sisense.com

2. How Do I Select The RightDatabase Technology?Your next step is selecting the technology configuration that suitsyour current business needs and will scale for future requirements.Are you going to use a data warehouse scale (‘big scale’) technology, or will data martscale technology be sufficient? To determine the database technology for yourbusiness analytics project, you will need to consider: the volume of your data and therate of increase, the number of users now and in the future, the number of datasources, the complexity of the data, if there’s a need for ETL, and the scale andscope of the project.Data Warehouse vs. Data Mart TechnologyDatabase technologies are used as the backbone for data warehouses as well as fordata marts. Since data warehouses typically store far more data than data marts, datawarehouse technologies focus on scale.On a single server, data warehouses are expectedto scale to terabytes of data (1 terabyte 1,000 gigabytes),while data mart technologies only work well in thegigabyte range.Other points of comparison are listed below:DM TechnologyDW TechnologyGigabytesTerabytes-PetabytesNumber of fields (1 mart)LimitedPractically unlimitedNumber of records (1 table)MillionsBillionsRaw data supported (1 mart)www.sisense.com

These differences also demonstrate why data warehouse technologies are often usedfor data marts, especially when the data marts are expected to grow in size. Theinverse is rarely true because of scale limitations. Using data mart technology for adata warehouse scale project generally won’t work as a long-term solution.Data Warehouse TechnologiesThere are three types of database technologies used for data warehousing:Software Appliances:A software application that might be combined with just enough operating system(JeOS) for it to run optimally on industry standard hardware (typically a server) or in avirtual machine. Examples of Software Appliances are: Microsoft SQL Server, MySQLand Sisense.Computer (Hardware) Appliances:A computer appliance is a separate and discrete hardware device with integratedsoftware (firmware) designed to provide a specific computing resource. Computerappliances are not designed to allow the customers to change the software, or toflexibly reconfigure the hardware. Examples of hardware appliances are: IBM Netezzaand Oracle ExaData.Distributed Databases:A distributed database may be stored in multiple computers, located in the samephysical location, or may be dispersed over a network of interconnected computers. Adistributed database system consists of loosely coupled sites that share no physicalcomponents (such as disk, RAM and CPU). Examples of distributed databases are:EMC GreenPlum, HP Vertica and Hadoop.The summary of differences between the different types of data warehousetechnologies is listed tedDatabasesCommodityProprietaryCommodity1 server1 serverUnlimited serversCapacityTerabytesTerabytesPetabytesHardware Costs4-5 figures6-7 figures5-6 figuresHardware ClassBest Architecturewww.sisense.com

Data Mart TechnologiesTwo primary data mart technologies are used today:OLAP (Online Analytical Processing):An OLAP Cube’s main purpose is to enable fast performing multi-dimensionalslicing and dicing. OLAP achieves fast performance by pre-calculating metrics (fieldaggregations) for all sets and subsets of unique values in all dimensions (fields)overnight. This avoids performing these slow operations in real-time during theworkday. Storing the results of these pre- calculations takes exponentially morestorage resources than the actual raw data does, limiting the size of raw data that canmake up a cube to gigabyte scale.IMDB (In-Memory Databases):Similar to OLAP, IMDB’s primary purpose is fast performance. It achieves fastperformance by loading the entire data mart into RAM, thus avoiding slow disk-reads(“I/O Bottlenecks”). The size of data mart is effectively limited by the size of RAM, whichtoday is limited to gigabytes in size.www.sisense.com

Which Technology Should You Choose?To determine the best database technology for your business analytics project, you’llneed to consider the number of data sources, the need for ETL, the complexity of thedata, and the scale and scope of the project.For projects with limited scope that utilize a single data source, a data mart solutionis probably your best bet. When your requirements grow to include multiple datasources with terabytes of data, along with your data analytics needs, a datawarehouse is the solution to support scale.Data Mart - When?When there is only a single datasource, which means the datadoesn’t need to be consolidated(ETL) prior to being delivered forbusiness analyticsWhen there aren’t many differentattributes and metrics tocross-reference (the data martdoesn’t need to have many fields)Data Warehouse - When?For a single centralized data store toserve multiple users and multiplebusiness scenarios (single version ofthe truth)When data volumes are large,rapidly growing or mayunpredictably spikeFor a one-time project (e.g. onedashboard), with no addedrequirements, new data sourcesor other changes expected inthe futurewww.sisense.com

3. What Do I Need To ConsiderWhen Selecting A Vendor?When you begin your research for a BI solution, you will quicklytake note of the sheer number of vendors out there. You mustnarrow the overcrowded field of vendors in order to come up withyour own BI short list.Exploring vendors’ websites is usually a great place to start your review. Based onwhat they share on their websites – and what they don’t – you can create acomparison map to quickly figure out which vendors are right for you and whichones probably aren’t. Armed with your short list, you can begin contacting vendorsto receive proposals.The Customers PageA BI vendor website without a dedicated page of featured customers is cause forconcern. Assuming the vendor does indeed have a ‘customers’ page, the first thingyou should look for is whether the featured customers are big corporations or smallercompanies.This is an important distinction because BI solutions for big corporations often havevery different functional requirements than BI for smaller companies have. Youshould focus on vendors that sell to companies similar to your own.If you need BI for an SMB, a BI vendor that lists only Fortune 500 corporations on its‘customers’ page probably won’t fit your needs as its solutions are probably too complex and/or expensive. Similarly, if you need BI for a large corporation and the business intelligence vendor lists only SMB customers, the solution may not deliver thefunctionality, performance or scalability you need.The second thing you should look for is whether you recognize any of the customerslisted on this page. Having recognizable names as clients says a lot about the credibility of the BI vendor. Well-known companies with recognizable brand names don’t trusttheir business operations to just anyone. A BI vendor with recognizable names on its‘customers’ page is less likely to disappoint than a company listing unknown names.www.sisense.com

Online Case StudiesReviewing the case studies on a vendor’s website is a good way to determine whetherthe solutions the vendor provides were implemented in business scenarios thatresemble your own. Credible vendors should include at least a few case studies ontheir website.It’s important to note whether the solution has been implemented in companieswithin the same industry as your own. Different industries will share many commonneeds, but there will also be special needs and requirements unique to each one.Another important thing to look for is whether the solution is being used in the sameway you’d like to be using your BI solution.Several types of BI solutions are available. Some focus on reporting, while othersfocus on analytics. Some are maintained primarily by IT, and some are designed formore hands-on use by business professionals. Case studies can help you determinewhether the solution can meet your goals and whether it can work for the intendedusers.The Partners PageThe ‘partners’ page is a great place to gather information about the vendor. You maythink that the ‘partners’ page is used mainly to find representatives in your ownregion, but there is a lot more you can tell about a vendor from this page.One thing you should look at is whether the vendor has a long list of serviceintegrators listed as partners. This is important because it says a lot about howdifficult the solution may be to implement. While software vendors prefer to sellsoftware licenses, service integrators typically make their revenue from projectsthey execute, charging by the hour or by the day.When a BI vendor has a long list of service integrators aspartners, it usually means the vendor’s solution is a challengeto implement and/or requires extensive, ongoing maintenanceand consulting. Otherwise, the service integrators wouldn’t seemuch business value from working with that vendor.www.sisense.com

The second thing you should look for on the ‘partners’ page is whether the vendor hasmultiple software/technology vendors listed as partners. Typical BI applicationsrequire several tools and technologies to be fully implemented, and when a vendorlists software/technology partners, it usually means the vendor provides only aportion of the BI stack itself.Download a Self-Service Free TrialEven though free trial downloads are commonly found in the software industry, theyare not commonly found on BI vendor websites. There are two major reasons for this:The first is that 99% of BI vendors are not software providers but rather serviceproviders. As opposed to selling products, they partner with existingsoftware/technology vendors to use those vendors’ products as building blocksfor implementing custom solutions.The second reason vendors don’t usually have a free trial available for download istheir lack of confidence in your ability to quickly and easily use their software.Generally these solutions are too complex or require professional services to getstarted.When a vendor does have a free trial version available fordownload, it typically means that the vendor is willing to selldirectly to you (thus removing the markups taken by third-partyservice providers and resellers) and that this vendor believesyou will be able to use its software yourself.The presence of a free trial also tends to indicate that the solution has shorterimplementation times without extensive third-party service provider fees.www.sisense.com

Professional ServicesVendors in the business intelligence space have two ways of making money: sellingsoftware and selling services such as consulting, implementation and maintenance.It is very difficult for a vendor to successfully focus on both, because they requirecompletely different core competencies, staff and operations. Building commercialsoftware in the BI space requires significantly deeper pockets than simply providingservices that utilize existing or third-party software. This is why the BI space has veryfew vendors that provide end-to-end ‘productized’ software and significantly moreservice providers that will custom-build a system.It’s important to understand that the total cost of owning a BI solution is rarely limitedto the cost of software licensing but, also, includes the cost of all the work that needsto be done in order to customize this software for your own needs.If a BI vendor presents itself as a software vendor but hasa wide range of professional services for sale, this shouldmake you wonder what it is about the vendor’s software thatrequires so many services.As previously mentioned, true software product vendors prefer to focus on sellinglicenses, not on selling services. It is very difficult to do both equally well, and from avendor’s business model perspective, it’s often wrong to even try. BI software vendorstry to make their software as easy to implement and as self-service as possible.Otherwise, they would have to spend more than they want on training, support andcustomization services.www.sisense.com

The Pricing PageA ‘pricing’ page is not something you often find on BI vendor websites. Those that dohave a ‘pricing’ page usually will not state actual figures but rather a prompt to contactthe company for a quote.The fact that a company doesn’t “name its price” on its website does not necessarilymean there’s a red flag. On the contrary—most BI solutions aren’t of the one size fitsall variety, as the required service and consequent costs to the service providers varygreatly according the amount of data that needs to be processed as well as itscomplexity.Additionally, some BI vendors do not publish pricing information because most of theactual cost of system deployment will go toward the implementation andcustomization projects. These projects are very hard to price before doing extensiveon-site analysis and estimating project scope. Therefore, standardized pricing wouldbe meaningless.Some vendors that actually sell software products will refrain from posting pricinginformation to hide the markups its service integrator partners get on the deal.In many cases, the service integrator is the one doing theselling to the end customer, and these partners tend toenjoy significant margins — another reason to steer clear ofvendors who don’t develop their software in-house.www.sisense.com

Online CommunityHaving an online community via blogs and forums is a very big benefit for customersin the BI space as it allows customers to leverage the experience of other users toaccomplish tasks and solve problems. Regardless of which vendor you choose, youwill always have questions, and it’s great to be able to get answers without having torely on the vendor’s official support channels alone.Online communities also give you a good indication of how popular the solution isand what types of things people are doing with it. Generally speaking, your experiencewith a BI solution will be much more fluid if you have access to the collectiveknowledge of an active online community.When a BI vendor does not have an online community, you should immediately beconcerned about how much support will cost you and how quickly you will be able toget responses to your inquiries.www.sisense.com

4. How Do I Launch A SuccessfulProof-Of-Concept?With your understanding of the software stack and differenttechnology platforms, you should develop a very short list ofvendors (no more than one or two) to launch a Proof-of-Concept(POC). These tips will put you on the right track for conductinga successful business intelligence POC.A POC That Doesn’t Use Your Data, Doesn’t Prove MuchFrom the vendor’s perspective, the most challenging part of any POC is getting familiarwith the buyer’s datasets and dealing with the unique challenges of that data.Therefore, most vendors will try to limit the scope of thePOC by reducing the scope of data sources involved,trimming the data down, using sample data sets instead ofyour own data, and so forth.Since dealing with your unique data challenges makes up 80% of the ongoingtechnical challenges you will be facing throughout your deployment, takingshortcuts here really means the vendor’s POC has proven nothing.Do Not Get Distracted By Pretty VisualsThe visualization of data is important, of course, but the biggest mistake buyers makeis to judge the vendor based on the pretty dashboard samples it shows you on itswebsite or during a demo. With visualization software components a dime a dozen,a vendor can easily ‘fake’ these pretty graphics.The real challenge is customizing these dashboards to your own needs and havingthem show your own data. This part can take most vendors months and can end upeating a large percentage of your budget. If a vendor can’t get your own data todisplay in dynamic and creative ways within just a few days, you should probably finda better one.www.sisense.com

Address Both Future & Present RequirementsMost vendors insist that you define your requirements very clearly to ensure thesuccess of a project of any scope, POC or otherwise. The catch is that it doesn't matterwho you are or how much experience you have — it is almost impossible to know inadvance what your future requirements will be.BI requirements tend to be highly dynamic because businesseschange all the time and business users are continually refiningand adjusting their requirements. Today’s reporting needs willlook very different in a year from now, and today’s analysis willlikely be relevant for only a short period of time beforebecoming obsolete.The best advice in cases like these is to focus the POC on the ability to meet ad hocrequirements, not just a set of fixed reports that might end up becoming irrelevant bythe time the project ends.Consult Your Own IT ProfessionalsIn many organizations business analytics solutions are already set up that are highlyreliant on IT. Business professionals, frustrated by not being able to extract relevantdata quickly and independently, will look for BI solutions that cut IT completely out ofthe loop, giving them the ability to manage their own analytics processesindependently.Nonetheless, it is still highly recommended that you consult your organization’s ITprofessionals regarding topics with which they are more familiar: scalability,integration cycles and so forth. Remember, the vendor has no real interest insharing issues that can trip up the success of any well-intentioned project.www.sisense.com

You Should Not Be Required to MakeSignificant Financial InvestmentSome vendors will promise you the world, but will demand significant upfrontinvestment to prepare the projects, hardware and software you need for a POC.Don’t agree to this. Demand at least one solid report or dashboard running over yourown data before you agree to any financial commitment.If a vendor is not willing to work with you before you put money down, it’s probablybecause that vendor would have to spend weeks on development before it can reachthat point. That typically means the vendor is either trying to sell you archaictechnology or simply trying to pull the wool over your eyes.Next Steps:Try Sisense for FREEJoin a Sisense analytics expert for a DemoContact uswww.sisense.com

Evaluate Business Analytics Software www.sisense.com The goal of business analytics and intelligence software is to help businesses access, analyze and visualize data, and then communicate those insights in meaningful dashboards and metrics. Un