Big Data Day LA 2016/ Use Case Driven track - BI is broken, Dave Fryer, Product Advocate, Domo

Editor's Notes

• #2 Personal intro, My name is Dave Fryer. I am a Product Advocate or Evangelist at Domo. It’s kind of a weird title but it essentially means that I get to travel around and share the vision of Domo with people. I think it’s a lot of fun and I really enjoy it because I think Domo is doing something to the BI space that should have happened a really long time ago. That’s kind of what we’re going to be talking about today… The object of today’s session is to point out what I believe to be 3 major problems that are currently plauging BI. I’ll also talk about some essentail characteristics that are present in organizations who are effectively using data to make decsions. Finally I’ll give you a brief demo of Domo. So let’s jump into it.
• #3 This is going to be review for a lot of you, but I want to define what I mean when I’m talking about BI in this session. When I refer to BI I’m referring to a traditional BI setup. I want to take you through the process of a traditional BI setups. Data undergoes a long journey before it’s ready for analysis. This overview will describe that process and system. First things first, you need data. And you have different types of data. A great example of External Events would be twitter or facebook. Someone tweets or posts about your company. External information and events that are effecting the perception of your brand, that lead to engagement and sales. Then you have internal systems. When I say business systems this includes things like your CRM where you house all of your customer and sales data, marketing platforms that you use to run marketing campaigns. All of the systems that help you track leads, how they’re converting to sales and driving the operation and success of your business. Salesforce, google analytics, omniture, netsuite, quickbooks. There are a ton of different systems here. Internal systems just refers to when a company sees fit to build their own application that helps run the business and generates data. This could be an internal wiki or customer relationship management tool. So these are all ways data can be generated but it isn’t ready for analysis yet. Maybe it is in little pockets – if you’re only concerned about sales data you can log into salesforce and get some numbers but if you want to look at the business holistically, this disparate systems don’t lend well to that. So traditionally what happens from here is that all of this data from all of these different sources need to live in one spot so the data can be commensurated and pulled together. This is often referred to as the back office. Again, these are kind of logical things. There are different ways or formats, but in general it’s the same process. Once it’s in the back office, it’s raw data. We need to prepare this data and transform that data to make it useful. This is where you business intelligence teams live, your data warehousing teams. Once it’s ready to be consumed to pushed over to the front office – your analysts and consumers. Once there we can start trying to answer questions, who are our target customers, what are some strategic decisions we’re making, etc. So that’s what traditional BI looks like. A big problem as I see it is that there are barriers between the business decsion maker, the end user, and the data. These barriers can be processes where you need to go through an analyst or admin to get data prepared, a lot of the time it’s user friendliness of the tools and systems. Business users simply don’t know how to use the tools.
• #4 It’s obvious if you look at the data that there are problems. 80% of BI projects fail (forrester). high expectations and aggressive timelines. regardless of the reason, there is an obvious problem with BI — business users aren’t able to use the data to make decisions. We’ve seen a lot of companies coming to us at Domo for help. A big financial services customer who spent in the neighborhood of $2M on an executive dashboard project. The project took about 18 months to implement only to be scrapped in 4 months. Why did that happen? Relevance. The project only included data that was accessible through their warehouse and the people trying to use the system were now having to spend more time trying to find relevance or correlation for the things they found in the system than they ever had been before.
• #5 When we talk about BI there are really 2 different audiences, right? There is the business side and then there is the IT side. This illustration highlights that there is a difference in perception, but the takeaway from both of them is that BI is broken. The things that really jumps out to me is that the majority of business users just think data isn’t available. The perception of failure gets plenty of spotlight from the business side — business users are always asking for “this”, this exact thing that I need to see. IT answers with “we have this. this is what we’ve collected and have available.” The problem is business questions don’t start and stop dependent on what data you have collected in your warehouse. IT feels the pain of failure because they are smart they do want to provide value. They’ve got a lot of systems that hold critical data. But it’s hard when typically step 1 is to pull the trigger on a 6 month ETL and warehouse migration project before they respond to the current business need.
• #6 And really it’s only getting worse. This is a graphic we released abaout a week ago. It shows much much data is being produced every minute in 2016. Massive amounts of data are being produced every minute both inside and outside the organization. Some of my favorites: Almost 1,000,000 tinder swipes 400 hours of new youtube content is uploaded. Amazon makes over $200,000 in sales And the big machine of the last year, 7 million snapchat videos It’s hard to make sure you’re looking at the right data but it’s only getting harder. Data never sleeps. New systems are popping up every day that provide data for your company.
• #7 So we’ve established that BI is broken. Business users and IT both think so, but why?
• #8 1: BI systems are not aligned with business drivers 70% (gartner) of BI initiatives product data that is out of sync with the business. People are reporting on things just because they can, not necessarily because they drive the business. Traditional BI systems are not designed for the end user. Unless you’re a SQL wizard, you’re not going to be able to have access to the data that lives in the warehouse And because your company signed off on the idea that unless the data has been processed and lives in the warehouse, it’s useless, time is now an issue. The concept of realtime or right time data is out the window. What are you going to do with yesterday’s weather today?
• #9 2: A focus on data collection, not on data use. BI departments become data hoarders instead of data users. A typical, traditional approach when companies start a BI project asks the question “where does the data live, and how can we get it into the warehouse?” Instead, they should be asking “what are some of your business initiatives? What problems are you trying to solve? What questions are you trying to answer?” From there the question can be asked, “where does the data live that will help us answer that question?” There really isn’t any point filling a warehouse with data if you’re not using it to answer business questions. FIX — Integrate, don’t migrate your data. Free your data. Put it in the hands of the people who use the data. In your typical BI project, 80% of the time and resources are focussed on data integration (forrester). The majority of the project should be focused on learning about customer behavior, eliminating guesswork, and finding new business opportunities.
• #10 3: Accuracy of the data. Bad data or poor data quality costs US businesses $600 billion annually. Data centers and warehouses are getting way too complex. After a while warehouses start to become a catch all for everything. With so many technologies creating a BI solution, pretty soon they become a mish mash of processes and data. Data is like a bacteria. When it is kept covered up somewhere, you really don’t know what you’re dealing with. You don’t know what is good, what is bad. What you really need to do is get it in the hands of not IT, but the people who are using the data to make decsisions. If there is inaccurate data, it won’t take long before the business user catches it.
• #11 Characteristics of good BI Data needs to be Accurate. Expose data to the business users through rich, interactive visualizations. No one can make sense of rows and columns of data. Data needs to be easily consumable so abnormalities can be investigated. Data needs to be timely. Remember, what am I going to do with yesterday’s weather today? Data needs to get to the end user, the decision maker in an appropriate timeframe. For marketers this may mean every 15 minutes. For finance this may mean every 2 weeks. The best way to do this is to integrate your reporting tool with your data sources. This way you’re cutting out the middle man of the front office, back office data warehouse. Have your platform act as the warehouse. Data needs to be actionable. The system needs to produce data that is relevant to the business. You’ve hired really smart people. If you serve up data in a way that is easily consumable you can help then make smarter decisiosn. If that easily consumable data allows for collaboration, alerting, and is accessable from anywhere, that’s where BI turns into business optimization. Timely Integrate, don’t migrate Actionable Collaboration, alerting, mobile
• #12 Set it and forget it.