Finding Clarity in the True Cloud

January 04, 2016

What would you consider the most overhyped tech term? Disruption? Big Data? The Internet of Things?

For our founder and CEO, Roman Stanek, that term is the “cloud.” Although GoodData was born in the cloud and we intend to stay there, we’re deeply invested in discerning the true cloud from “cloud-washing” of older technologies. To many, this may just seem like a marketing ploy--a trick of language to show how GoodData rises above the rest--but it has real impact on how business intelligence is implemented, maintained, distributed, and ultimately, monetized. And as cloud adoption rates rise to 44% this year, and anticipated over three-quarters in 2016, defining cloud technologies is an issue that will affect most of the market.

“Traditional approaches to business intelligence (BI) delivery 
frequently fail to meet business needs in terms of agility and speed. 
As a result, business and technology professionals are increasingly 
looking to cloud deployment options to obtain the required BI 
capabilities in a more flexible manner.”

-The Forrester Wave™: Cloud Business Intelligence Platforms, Q4 2015

From security to measuring user engagement, true multi-tenancy in the cloud has clear benefits over cloud-washed technologies of old, which is why we’ve hung our hat on true cloud business intelligence.  We are in unusual era in the cloud market now, where transitional technologies appear to be the more comfortable choice-yet prove only to perpetuate the limitations of old architectures.  Very few vendors can talk about delivering systems of insight, and then actually prove their case.  We can, and it all boils down to the merits of our deployment architecture. To illustrate, let’s explore the differences between cloud and cloud-washed, and how they can affect your business – this new DataTalk video provides some added perspective:

Avoid Strange Clouds: 5 Signs of Cloud Washing

Our security advisor Bil Harmer has talked extensively about the implications of “cloud washing,” which paints solutions that resemble older, on-premise architectures as “new,” despite a lack of true multi-tenancy. However, business intelligence users may not always recognize these strange clouds right away, which is why I’ve created this field manual to spot old, hosted models masquerading as the multi-tenant cloud.

Warning Sign #1: Desktop instances & offline mobile analytics.

Software-as-a-service apps like LinkedIn and Gmail require constant internet access, and it’s okay for your cloud business intelligence solution to require it as well. Any supposedly “SaaS” or “cloud” application that touts desktop instances and mobile apps that don’t require an internet connection are actually software, without the as-a-service. The likelihood that your devices are offline for any extended amount of time is shrinking, not growing; so just bring a browser.  Further, almost all of your planes, trains and busses are online, so if you’re out in the woods, bring a PDF and a coat. This is a novelty feature, period.

Warning Sign #2: No possibility of benchmarking.

Another hallmark of the old, hosted model is a lack of benchmarking capabilities. If your analytics deployment is truly “one to many,” that means your cloud framework individually secures each organization’s data while also being able to evaluate one against the mean--thereby building a performance benchmark index of different franchises, suppliers, partners, and end users. If your business intelligence vendor says “only if all the data is stored in one instance,” when you ask about benchmarking, you should question their multi-tenancy, immediately.  If they store individual organization’s data in the same operating instance, then you risk exposing the data for all organizations if a single user account is compromised.  True clouds share resources, but secure data individually.  Ask Salesforce how they do it.   

Warning Sign #3: The need to query on-premise data.

The discussion of analytics and business intelligence platforms has long revolved around siloes, and yet many platforms rely on querying siloed, on-premise data as part of their “cloud” offering. This emphasis on gradually moving organizations to the cloud in a cloud-first world is hurting the practice of BI. The reticence to load data into a cloud environment has very little merit.  A large data load is only challenging the first time, and loading any data, even aggregate, has the benefit of long-term persistence and reuse as a snapshot or benchmark. If your concern here is security, then you need to remember that your organization’s most sensitive data--payroll, taxes, sales forecast and orders are already in the cloud if you use ADP, Salesforce or Netsuite. If your “cloud” vendor is touting their on-prem query prowess, they are hurting your progress not helping.  There are much larger benefits if you centralize your on-prem, SaaS, Social and IoT in a true multi-tenant cloud.

Warning Sign #4: A lack of platform usage data.

In the consumer SaaS application space, real-time usage analytics have become the norm. Facebook has Facebook Insights, FitBit has Dashboard, AirBnB has Analytics. For purpose-built cloud applications, these usage statistics shouldn’t be hard to find or provide, since all of your users live on the same computing fabric, and tracking and sharing their actions within the platform should be easy. That’s why usage dashboards have become par for the course in cloud applications, and like benchmarking, should be one of the first offerings your cloud BI vendor mentions when they talk about their platform. If they cannot provide that information, such as “Daily Average Usage” (DAU) rate, then it means that they are very likely managing siloed instances of their product and can’t measure the activity rate of their entire installed base.  So, they gain no objective insight into the behavior of their community. Again, ot a true cloud.

Warning Sign #5: Multiple codebases.

While this is the root cause of all warning signs above, it’s also the easiest question to ask when evaluating cloud BI vendors--Do you have one codebase? As Bil pointed out so poignantly in his blog post on the subject, the simplest definition of the true cloud is “a multi-tenant SaaS deployment, which has a single binary or deployment used to serve multiple customers. The unique element of multi-tenant cloud vendors offer is the shared codebase, which is rendered on login and able to interpret configurations to create a unique view for each customer.” The advantage is that we update the environment once and all customers are up to date, which is especially useful when “Zero Day” events occur and you WANT to be current.  The cloud isn’t easy to breach, and you shouldn’t need to decide when it’s time to upgrade or re-secure your environment.  It should be automatic and it is if you’re true cloud.

Bonus!  Warning Sign #6: User licenses.

Multi-tenant cloud systems should encourage user adoption, not thwart it.  Vendors who license on a per-user basis, are stuck in the old world.  

Cloudy With a Chance of Progress: Realizing True Cloud BI

I have long believed that business intelligence needs the cloud, as much as other software-as-a-service platforms need the features the cloud provides. And at long last, influencers in the analytics and BI space have come to agree with me, focusing on the completeness and usability of different platforms, rather than their ability to “play nice” with older, hosted systems.

“It’s not the individual capabilities that differentiate the Leaders, 
but rather the completeness, comprehensiveness, and integration 
of the entire suite of agile capabilities that sets the Leaders apart 
from the rest of the vendors in this category.”

-The Forrester Wave™: Agile Business Intelligence Platforms, Q3 2015

The rules of “success” in cloud analytics have changed, since less than one third of organizational data is actually used for analysis, and ultimately, insight. BI is no longer a game of data collection and storage, but a system to get to actionable business insights and outcomes across an organization, in real time. As Roman has often said, Big Data is no longer about storage and access, it’s about creating value for every customer, every partner, every user. As more BI platforms migrate to the cloud, it will become imperative to purpose-build analytics in the cloud, instead of cloud-washing old, hosted, on-premise systems and appliances. This is the only way we can finally utilize the full suite of benefits the cloud provides, and the only way we can progress.


Want to ask about something specific?

Contact us