Each year I look at Forrester’s predictions because they have a great understanding of the market. Their research approach is product-centric. They’re hands-on with the product and understand where it’s going rather than just relying on customer surveys.
This year Forrester has made some good predictions for BI that have got me thinking. I wanted to take the opportunity to reflect on where I agree (and disagree) with Forrester’s predictions for 2018.
Agree: AI will have a meaningful impact on BI
There’s a lot of hype around AI and what it can do in the BI space. Forrester has predicted that the vast majority of vendors will start to use AI to automate their products. They see that AI is crucial to automating insights and helping people get information faster.
From our perspective, we’re seeing that many organizations have acquired AI technology and the capability to create predictive models. I agree with Forrester that we will see a huge shift towards customers leveraging AI to automate their analytics.
Agree: Forrester is right, data lakes will die
In the big data space, people have been excited by the ability to store lots of data. They haven’t worried too much about what they should keep and what they shouldn’t, so all their data has been thrown into a data lake. The actual result has been to create data swamps.
It’s very hard to pull data out of these lakes if it’s not structured in a way that’s usable. If you don’t know it’s there or how it’s joined, the data loses its context and becomes worthless. That’s why I agree with Forrester that data lakes will die.
Organizations will literally need to throw out their data because there’s no compelling reason to keep everything forever in the hope that maybe you’ll use it one day. Over the coming year, organizations will start to clean their data swamps. Many will just switch off the infrastructure because it’s no longer relevant to their business.
Agree: Organizations will re-centralize BI
Organizations tend to go through cycles where they push functions to the frontline and then bring them back centrally. In BI, there’s been a medium-term trend for analytics to be delivered in a decentralized model. Practically this means that organizations have less oversight or governance over their data and this increases the propensity for errors in their analysis.
Chief Data Officers are now expressing a desire to bring this back in centrally. That’s why I agree with Forrester, that we’ll see organizations wanting to bring BI back towards central control. More organizations will develop BI centres of competency, especially if they can leverage tools that allow them to get insights back out to the business quickly. If CDOs can deliver the ‘why’ faster then centralised BI will be here to stay.
Disagree: Conversational interfaces won’t have the impact Forrester predicts
There has been a lot of discussion about conversational interfaces like bots. While these are an interesting concept, I don’t think we have the technology to deliver a truly viable solution for users.
Conversational interfaces simply aren’t efficient enough. While a user may be happy to type a question for a simple query, it’s not as easy when the problem is complex. If you want to analyze particular product sets for a particular country for particular a time period then it’s probably quicker for someone to click rather than type.
Command-line interface died because there were more efficient ways for people to interact with computers. Using a conversational interface is taking a step backward in terms of technology. If you can get an answer in one or two clicks why would you want to type a sentence or two?
This is really a question of efficiency. Vendors should be creating products that are as efficient for users as possible, whether that’s through clicks or conversational interfaces. So, the real question is whether vendors can optimize their interfaces for efficiency – my bet is that conversational interfaces will lose that battle.
Disagree: The future of BI is not in the cloud
Forrester, and many other people, are saying that the future of BI is in the cloud. While there are some great cloud vendors out there, and Yellowfin does deploy in the cloud, the cloud is not the best solution for everyone. What’s best for an organization comes down to data gravity.
Your analysis should take place wherever your data is. The vast majority of organizations keep their data in-house, not in the cloud. It’s very inefficient to have a cloud BI product if all your data is in another place. So, for the moment and the foreseeable future, I think we’ll see on-premise remain the primary delivery model for BI solutions.
This year is shaping up to be a time of change in the BI space, only time will tell whether Forrester, me or both of us got it right.