Digital analytics role is not just defined by web analytics or analysts

My previous blog post talked about treating data as an asset, and would like to focus on the human resource in this post.

Roles and Responsibilities to turn data into an asset

Many companies still think hiring that one web analytics analyst is a great investments. Most of these companies forget that digital analytics discipline is getting more deeper and wider (or complex). I strongly agree with Avinash’s 90-10 rule where 90% should be about people/analyst than tools where it should be 10%. This post is not about that, but focusing on that 90% and really understand why we need smart engineers and analysts behind digital data. Businesses who are new into web analytics need to start thinking about analytics from end to end planning, and NOT just from hiring someone who can look at data coming out of Google Analytics, Webtrends, SiteCatalysts, etc.

What are the common digital analytics disciplines and practices companies should think about (let’s assume these companies are new to this…)

Digital analyst’s disciplines and potential expertises required

• Web Site Analytics (including eCommerce marketing)
• Media planning and buys (including but not limited to SEM, eMail, Banner Ads)
• CRM
• Social Media Monitoring
• Testings: A/B or MVT
• Modeling (marketing mix, forecasting or predictive modeling)
• VOC (voice of customer) or Consumer Insights
• Market analytics or Competitive insights

Imagine the engineering side of these data sources. The engineering side needs following accumen to help turn these data sources into readable data for these analysts.

• Data integration: engineering skills required to bring various data into company’s environment.
• Translation of data: Bad data in, Bad data out. You can still have “Good Data In and Bad Data Out”, if the folks who are working on the logical layer messing up the translation after data comes into database. This a critical piece to make sure analysts and businesses have good data to analyze.
• QA and governance: experts who understand the business needs and making sure the implemented solutions are accurate, not breaking things, and checking to make sure business has good data on hand.
• Expertise in infrastructure: Yes, all these stuff around data needs to be processed, stored, and reported. Where and how? Engineers need to understand the technical environment surrounding the data, and someone needs to own maintaining it.

Even if you’re not thinking about one head count for all of these disciplines, let’s look at where these data come from and ask your self how much level of expertises or resources are needed to “manage” all of these.

Data Sources and potential role name

Web Site Analytics // Web Analytics Implementation Engineer: Data typically comes from analytics platform’s given javascripts embedded on the website, and the analytics service provider would process the received data within a day to aggregate necessary data to report on visitors, sessions, page views, and tie all of these to user behaviors or action events. These could be implemented by web engineers, but in most of the case pure engineers would not understand what tracker means what to business. Dedicated implementation engineers usually requires a strong business acumen so they can connect the dots between javascript to business needs.

Media Planners & Buyers: Web Analytics data captures great post click data, but not on the buy side. There are many companies specialize and are dedicated to the planning, buying, and executing the ads. In digital space, these data are planned in Ad Serve tools that aren’t web analytics (obviously…). In addition, recently, with YouTube gaining momentum replacing TV in gaining eyeballs at low cost, data around these platforms are becoming important as well. YoutTube impressions are replacing the TV GRP’s (j/k).

CRM // CRM manager or analysts or CRM data engineer: In most cases CRM data are stored in a form of database. In some cases those data in house are accessible by 3rd party depending on the company’s contract/relations with the agency. In other cases, the database is hosted and managed by third party service providers (sounds like a bad idea.. but you’ll hear it time to time). Data needs be pushed into the database from some where. These managers should be the expert in the data in/out aspect as well as working with analysts to understand the customer profile by profiling the data. This discipline has two required aspects, both technical side as well as the management side. That is why it is really hard to find one person with both talent, because database engineering and managements are two different skill sets.

Social Media Monitoring // Social Media Analysts or Operations Engineer: Most of the social media monitoring enables the data through paid services or platforms (like Radian 6, CrimsonHexagon, etc.) to collect certain phrases people are talking about and less about tagging. However, from data stand point, you’ll need operations engineer to bring the data into the data warehouse. Otherwise, you’ll be sitting on 3rd party tool collecting and hosting the data for you. Good luck to you when that company goes out of business.

Anyways, Social Data has many text inputs from the users, and to mine those data, it will be challenging and hard to bring in all those data into a data warehouse. But not a problem in an era where Big Data is a commodity right? Well, think about it, you still need data in house if you really want to treat Social Data as your asset.

A/B test analyst // Test Planner // Test Creator: Any web site testing involves planning, and that means you need to have a good planner. Once great plan is locked in or approved, assuming all creative assets are final, then you’ll need to have a technical person who can enable the test by setting it up the event tracker on site. Then test needs to be set up in the Testing Tool as well.

This is a lot of work, and planning and engineering are two different skill sets. If you really want to test multiple times in a month, then business really needs a strong commitment on resources. Learnings from test and the culture built will need to be an asset for the company, so make sure to have those knowledge and learnings stay (perhaps in intranet, or corporate’s wiki site, etc.).

VOC (or Voice of Customer) // Consumer Insights: Quantitative data from analytics alone gives you an insight if you synthesize massive amounts of data to that one learning. However, you can not ignore or argue with what your customer directly tells you. When that VOC data marries with quant analytics, then it is very insightful. That means the data sources would come from these survey tracker implemented on the website, while making sure you have a good designed survey questions to answer critical business questions. Some one needs to be able to plan the end to end aspect of the discipline. One analyst can not do that especially if you’re going to take VOC to a new/next level to best learn your customer’s pain points.

Market Analytics Analysts or Competitive insights: Typically these data sources come from 3rd party paid services like ComScore or Hitwise. You may think, no heavy requirements needed, sounds easy. Yes if you end there then web analyst could take on the rest. However, if you’re in a business like buying data from POS (point of sales) or analyzing market share, then it is a whole another level of man power needed to make sense of it. Web site analyst looks at owned media alone may be missing the big picture when market share of your company is increasing/decreasing by 2x. If you ever seen those data directly from NPD or GFK, and not just Compete.com stuff, then you’re going to say ‘wow’ I need more time to analyze all these data and see how that is impacting consumer behavior online.

My intent of this article is make sure analysts are thinking about the data beyond your one segment of your digital discipline. Data experts are involved with many experts outside of marketers and they are all from data where it starts from some where. Managements who got Web Analytics 1.0 to 2.0, now needs to think about data as an asset, governance, privacy, security, data warehousing, etc. Not to mention growing number of data and sources CMO thinks marketers already have, so take action before business demands it. Start by looking at the resource on hand and the reality of the data eco system.