TL;dr
How do you get the decision makers in IT to take analytics seriously? Members of the Digital Analytics Association come to the aid of one of their own with sage advice including:

Data Solution Architect within the IT department
Grow your internal education offerings
Combine tag management with data prep, discovery and discovery
Make IT feel they are helping you rather than taking orders
Have senior managers the value of analytics to the business
Mark unfinished tasks with an “awaiting IT development” flag
Share the results and the credit with IT
And finally, a detailed to-do list that make Scrum/Agile work for you

Stop Making Analytics an Afterthought

The Digital Analytics Association has taken a beating in the past for “hiding value behind closed doors.” This sort of grumbling generally comes from people who run their own, small agencies and simply cannot fathom why somebody should have to pay $199 a year for stuff that should be free. The short answer is that stuff costs money and when you get a lot of great stuff together in one place, it costs about as much as it take to support a two-Starbucks-per-day habit for a month.

DAA

I bring this up to express the administrative danger I’m going to be in when this post hits the tweets. I’m about to shine a light on one of those secret, eyes-only things the DAA has stashed behind the velvet rope.

Why risk having six black Escalades to pull up in front of my house, leaving me with only a moment to tell my wife I love her before starting my new life in Guantanamo? Because the following is so valuable and gives a taste of what the DAA is selfishly hiding from those who can’t cover the annual equivalent of office supplies for one employee.

Laura Asks for Help

Laura Reed is a Digital Intelligence Engineer on the Adobe Analytics Center for Excellence Team at  Vanguard. She’s no lightweight. Laura has served in a variety of digital marketing and strategy roles for such varied companies as GfK North America, Digitas, Bristol Myers Squibb, and Shire. In addition, she co-founded a startup focused on building a consumer electronics product with an associated location-based mobile service that received funding from Liberty Media. She has also presented at a DAA Symposium on a panel on The Art and Science of Evolving Your Career in Digital Analytics along with other masterminds Randy Zwitch, Senior Data Scientist at Comcast and Quentin Jones from Barclaycard.

Even with her significant experience, Laura came across a problem that she recognized required more experience than she had, so she reached out and asked:

How Do You Influence IT and the Business to Stop Making Analytics an Afterthought

ITI’m currently working on a project that is designed to influence IT to make analytics a key component of their process, as it is often an afterthought today.  I’m interested to see what techniques others have used to do this successfully.  Is this still a problem in most organizations?

Oh boy, is it ever.

Sushant Ajmani, Head of Consulting at Nabler in Bangalore was the first to come through. He acknowledged the universality of Laura’s question and jumped at the chance to preview a blog post he was working on at the time about setting up Data Governance to maintain quality and integrity of online data. Clearly, IT plays an imperative role here.

First and foremost, it’s a Universal problem in our industry, and I have seen this with more than 200 companies I have consulted in last 15 years but, in last 3-4 years, there has been a significant improvement happened on the IT front, and lot of engineers in my network are taking active interest in learning about the digital ecosystem and putting a lot of efforts in learning about Implementation Audits, Tag Management, Constructing a Universal Data Layer and most importantly, how to prepare data for Visualization and Exploration Analysis.

Secondly, Analytics has always been an adopted baby of the Marketing department despite of the fact that; it’s highly influenced by the IT department and requires constant governance from IT point of view to maintain the quality and integrity of the tags. This current structure has to change and the best way to tackle this problem in the short term is to create a specialized role of Data Solution Architect within the IT department. A Data Solution Architect is an individual who comes with the following back-ground and possess the following skills:

  • An Engineering Graduate
  • Analytical
  • Strong Problem Solving Skills
  • Extremely passionate about learning the architecture, underlying components and leveraging the out of the box packages
  • Spent at-least 5-6 years in the Digital Analytics domain and worked closely with the Marketing and Business Teams
  • Able to translate the business needs in to technical specifications
  • Extremely good in creating a Digital Roadmap by respecting the organizational IT and Digital infrastructure
  • Authoritative
  • Excellent Communication and a nice story-teller
  • Ready to take risks and fail faster

This individual plays a liaison between Marketing and IT team and maintain a lower barrier of entry with very limited friction.

Thirdly, in order to empower your IT team to handle the data respectfully and maintain integrity, it would require series of LEARN and EARN sessions in the company, and here are few sessions I would highly recommend for your IT Team:

  • Creating a Universal Data Layer
  • Tag Deployment using Tag Management Solution
  • Tag Auditing
  • Data Governance
  • Data Interpretation (This is extremely important to maintain the data quality and integrity)
  • Mobile App Tracking
  • Data Integration
  • Data Preparation for Visualization (Tableau, QlikView)
  • Data Preparation for Advanced Analysis (R, Python)

To get the best outcome from these sessions, make sure the sessions are conducted by the in-house Solution Architect or, an external consultant, who have excellent technical background and can empathize with the current state of your IT Team and could put him/herself in their shoes. This is extremely important. And please, don’t forget to structure these sessions in such a way that there is an incentive behind learning and people feel motivated to learn.

Fourthly, the IT Team gets bored pretty fast if, you ask them to do the same monotonous activity of maintaining and optimizing the tags using Tag Management Solution because after some time, it becomes damn boring and people start doing silly mistakes which indirectly impacts the data integrity so; to maintain their momentum and interest, allow the following techniques:

  • Combine the Tag Management work with Data Preparation and Exploration work. Ask your IT individuals to learn the basics of R and Python and develop their Data Science skills. They will feel very motivated to work on this because, there is a lot of buzz in our ecosystem about Data Science.
  • Combine the Tag Management work with Data Visualization work. Ask your IT individuals to learn Tableau and Qlikview and build smart visualizations for the business, marketing and executive teams. This will give your IT team a lot of visibility and accolades and motivate them like anything:-)

Lastly, here are few process level changes I would recommend:

  • Every time the business team submits a request for implementing a new feature on the website or, change the page layout or, roll-out a new widget; ask them to provide a single metric and dimension that they would like to track against this requested change. Create a Project Request template in which, ANALYTICS SECTION should be mandatory to fill.
  • Follow a templatized approach for tagging your pages. Every time a new page gets launched, the following bare minimal tags should always be there at the template level irrespective of whatever analytics tool you guys use:
    • Page Name
    • Page URL
    • Page Type
    • Site Section
  • Create a solid regression testing suite to quickly identify the Tag Integrity issues during development. To accomplish this, you can either use a standard testing tool or auditing tools like Observepoint. These days, most of the enterprise tag management solutions also offer preview mode to quickly do the sanity check of your tags.
  • Introduce Tags and Reports validation a part of your QA cycle and make sure, no code will go to production unless and until it has been validated by the QA Team from the Tag Quality and integrity perspective.
  • Post Launch, setup your Automated Tag Validation Simulations through tools like Observepoint and gauge the quality and integrity of your tags in a proactive manner

Sushant wraps up with a common offer in the digital analytics community:

I hope the above suggestions would help you to empower your IT team, and in case you have any follow-up questions, please feel free to revert on this network.

But the DAA Community Forum wasn’t content to just leave it at that. The inimitable Stéphane Hamel, he of the ground-breaking  Digital Analytics Maturity Model, piped up as well.

Wow! The answer from Sushant Ajmani is just awesome and to the point.

IT2Misunderstanding between marketing and IT doesn’t really need any further intro. When speaking at events, most of the attendees are coming from a marketing back  ground and they naturally tend to see the world through those lenses. I’ve witnessed marketers despise the role of IT as merely being “doers” of their wishes and commands. The folklore is also strong around pizza-eating, overweight, thick glasses nerd in his basement (how many time I have seen this “computer nerd” picture in presentations!)

Typically, the primary role of IT people is to be “problem solvers” – leveraging the right technology to enhance and optimize business processes. The worst thing to do is to come to them with “orders” rather than extending an hand and saying “I have this problem and I thought you might help”. Make your IT team be just that: a team! Involve them in the process – you might think you have the solution, but the reality is your IT people might have a much better one!

Probably similar to what Sushant described, I have spent countless time helping organizations (and now agencies) grow their digital analytics maturity by getting the team around the table and uncovering the strengths and weaknesses in 5 key areas of success: governance, objectives, expertise, methodology and technology. This maturity assessment is merely the spark that can lead to a more realistic digital roadmap and better understanding of everyone’s role & responsibilities. I sometimes do what I call a “RAS workshop” (Responsible, Approve, Support): we list tasks and assign responsibilities, and in doing so, develop better collaboration, team building, and just plain and simple clarity.

As if that weren’t enough, Vinay Jagannath, Digital Strategist at Nabler reminded us:

The Main pain point from IT’s perspective here seems to be that any Analytics especially Analytic consultants are external meddlers who are going to increase the complexity of work without benefits and <insert any consultant or MBA stereotype here>. The chief issue here is that IT is usually isolated from the business and agnostic to achievement or non-achievement of business objectives. One good Idea could be to get senior stakeholders (with IT and Business responsibilities) involved in demonstrating to IT the specific benefits that Analytics brings to them. Of course, in absence of access to such Stakeholders (if the Client company is very Large in size), it falls on us as experts coming in to do the bulk of communication.

To which, a (declined to be named) Digital Analytics Manager said he felt Laura’s frustration and suggested Scrum was partially to blame and suggests prioritizing the backlog for the IT team.

I believe the scrum model that IT has adopted over years has only made this worse, as often times Analytics is a piece of the puzzle/project that isn’t required to go live and gets pushed into a “fix it later” bucket of work.  Problem being that the “fix it later” bucket never gets picked up in sprint cycles and is prioritized behind new features or bugs.  Only when executives yell for metrics is when they will pay attention to that bucket.  What even makes it more infuriating is that the Analytics piece isn’t too complex to a developer.

Co-sign with Sushant that having a role within the IT umbrella is the best solution but is a longer term solution.

In the short term…

I have a little success trying to make part of the project managers workflows include Analytics and try to get them to have a mindset of a project doesn’t get pushed until Analytics has signed off.  Guessing your org. is similar any other IT department and has a backlog of 10,000 things to do.  Part of the tradeoff I used with PM’s is that I would help them prioritize their backlog (i.e. that bug on the list is only impacting 0.1% of traffic or that new feature will touch XYZ% of traffic and drive $XYZ revenue).

I have also had success by starting to include “awaiting IT development” in reporting to executives.  After a couple months of having this glaring line item in there reporting they eventually asked the IT executive what was taking so long.

Next came that Gilligan-of-Yore, Tim Wilson who weighed in with a long-term play that has served him well in the past: sharing the credit.

Share the results of analyses with IT. This can be a little chicken-and-egg, in that, if the most promising analyses were stymied by unimplemented tagging customizations, then there aren’t necessarily great analyses to share. But, when you do get data collection implemented in ways that you’re then able to get meaningful information out, close the loop with IT. Show them the results that the business got really excited about, credit them for enabling that to happen, and see what ideas the analyses spur for them. That really makes it tangible why, although tagging is not actually a visible part of the customer experience, it can be extremely valuable.

I’ve had that approach generate staunch advocates for analytics within IT who are in a lot more meetings and much more embedded in development processes than I am.

Last but far from least is that Twitter master, Jonathan Longden, Head of Digital Analytics & Optimisation at Sky (not  a spelling error – he’s British!) (His worthy Twitter account):

Having set up teams and operations many times myself, I have always found 2 ultimate problems:

> Nobody other than an analyst will ever *truly* care about analytics

> IT/technology/developers hate being told what to do by a completely different department/team

Therefore, I believe the only real way to get around this is to remove the separation that exists between these 2 areas. Put the analyst in proximity of the technology teams so that they can be properly pro-active and get things done; make the technology teams ‘analytical’ by incorporating analytics resource into the fabric of how they work; include the entire team in the development of hypotheses so that they don’t feel dictated to.

I have delivered this at Sky by embedding my team into the Agile development squads. Our digital properties (sites, apps etc., but also channels e.g. live chat and so on) are split into numerous product areas, autonomously managed by scrum teams, working Agile, containing: the product owner, digital optimisation analyst (my team), UX, developers, and so on.

Sprint

This is founded on several principles which I protect vehemently to make it work.

My team are:

  1. Product specialists – they do everything (tagging, data, analysis, testing, optimisation etc etc) for a product area. There is therefore no waterfall process to any of analytics and optimisation, and no need to use anyone external.
  1. Embedded – they are physically located in the development squads and are part of the team. They participate in all stand up meetings. They work directly with the BA and product owner to drive the product roadmap/backlog etc.
  1. Lean – they do not spend 3 weeks making PowerPoint and then doing a big reveal. They deliver their insights in standup format and feed hypotheses into a continuous improvement plan which feeds testing. They don’t do ‘reporting’.
  1. Pro-active. As aforementioned, IMHO nobody other than an analyst will ever really care about analytics in the same way. You cannot always expect people will ‘brief’ you with requirements for analytics which genuinely make sense. Therefore, the analyst is pro-active at delivering KPI improvement and doing whatever it takes to make this happen; they conceive and drive the work.
  1. Business funded. Their time does not need to be assigned to project codes, which is a sure-fire way of ensuring no analytics ever gets done, as it is the first thing to get dropped when budgets are tight, and they always are. They are 100% funded by the business area and free to practice their craft.

This is working really well for us and has been very successful in solving the problems discussed here.

All of the above is just such sound and in-depth advice that I felt I was too late to the game to offer any additional insights for solving this problem. But I’ll never be without a few key concepts when this topic rolls around again.

Thank you for a toothsome question, Laura,  and to DAA members for being community-minded supporters!

Share This