Responsibility – best served in layers

Imagine I would design and build footballs.

I mean footballs as in “football”, not “American Football”.

This distinction is pretty much where my knowledge of and experience with football ends. I played 6-a-side in the UK for 10 years and scored exactly 10 goals in that time. I am absolutely rubbish. I think I managed to be a nuisance in defense, but that might just be wishful thinking. I also don’t enjoy watching football matches, have no favourite team, and don’t even care who wins.

Now imagine I was asked to design and produce balls. For the world cup, no less!

How good would those balls be? What about the world cup? Players and fans puzzled or upset, totally random results, nobody would be happy. The press would go mad about the balls, the person responsible (me!), and whoever gave me the job. Carnage!

And yet.

Everything is possible in the land of digital

Working in technology, especially software, or marketing, you have likely seen things like this happen.

It is entirely possible, nay, common that IT “manages the deployment” of an Analytics or Targeting solution, or any other Marketing tool.

As you know, there is no implementation phase for Analytics, and A/B testing or Targeting should probably not be driven by IT, either.

I’m not saying they can’t do it. The goals simply do not align.

Analytics and Target are incredible with an ad-hoc approach, a relentless drive for incremental change, so to speak. Analytics deployments are not a project, they always evolve. Target, in its very soul, wants to change the site, constantly.

There is risk involved with both! Both tools are often in the hand of people who take risks. They break stuff, sometimes. And that, in our eyes, is perfectly ok!

And so, if you ask me, the right people to build that ball are the ball makers, and the right people to deploy, manage and use Experience Cloud solutions are Analytics / Testing / Marketing people.

To be totally honest, the comparison is a little bit unfair.

The obvious parallel would have been some small business owner building the web site for their venture, and using blink a lot. That can work, but most often, it does not.

I have seen IT or dev deploy Analytics and Target. It’s not as bad as a bad site, but it rarely leads to value or happiness.


So if someone in the Marketing org handles Target, who is ultimately responsible for the web site? And don’t you dare say “shared responsibility”!

The idea is that there can be layers, and specific people or teams who are responsible for some of those.

I still think that the combination of a data layer and a Tag Management System is an awesome opportunity to define layers. I still think that at this boundary, the responsibility of IT/Dev ends, and Marketing takes over.

I also still think that the boundary or interface must be well-defined, and under strict change management. This helps IT/Dev just as much as it helps Marketing.

Since the ones I have to convince to put the boundary under strict change management are likely Marketing: think about how easy your job will be once you have a solid foundation! Think how much you’ll be able to do!

And the other way round: Apple (to pick the blatantly obvious example) designs their products, then lets them build by someone else. The designers are responsible for what people see, touch, feel, and so on, while manufacturing is responsible for making it.

In our industry, we’re trying to do the same, but I think we’re not yet really good at it. We have a lot of (sometimes really bad) craftsmanship, and sometimes manufacturing doing the design. Both extremes are suboptimal.

My feeling is that things can start small, and I hope that dividing the work and responsibilities between dev and marketing might be such a start.

Digression – Org Structure

For this to work, Analytics and Marketing have to bring the right skills.

(4 years ago, I wrote an article about that in German — Mindestens Zwei, which is still relevant as of July 2019, so I translated it: Analytics Team Roles)

In this context, the important skill is Javascript and a certain affinity or at least tolerance towards coding and tinkering.

Thing is, in our trade, that is not necessarily a primary feature. People in Analytics love data, statistics, visualisation, or maybe politics. But a lot of companies have someone who, our of sheer necessity, started to learn how to deal with the code aspect. Often, this person drives everything, at least for a while.

This is the first level or org structure when it comes to Analytics, Testing, and Digital Marketing. This is how companies, big and small, start out. Some people call this level or model “lone wolf”.

Once there is some momentum, most people switch to some variant of the “hub and spoke” model, i.e. a (central) team handles core activities, while trying to get stakeholders everywhere to self-serve as much as possible.

There were a bunch of sessions at the Adobe Summit in London on that topic, and I really liked how this year, they went beyond “we are planning this”, and talked about “we did this, and then that happened”.

(If you don’t mind typing your email address into a form, you can watch the sessions online: From Reporting to Automation: How a data-led mindset can change your organisation, A data democratisation journey: From ticket queue to self-service)

Anyway, once your hub and spoke is a well-oiled machine, you might want to go to the next level, to what most people call “fully integrated”. We think of an organisation in which every single function uses data, relevant to them, to make decisions, where everyone knows both bigger picture and where their part sits.

Don’t be jealous. I don’t think anyone is actually doing that right now. I personally only know one single company who even tries to get there, at least here in EMEA.


Most people who do Analytics struggle with making a big impact. There are a lot of factors why that is so: data quality, organisational resistance, missing resources, and many more.

In the context of this article, I think data overload and aimlessness are biggies. People “measuring everything” are still the majority, even though we all keep talking about “one thing at a time”, “keeping it simple”, or “concentrating on the real issues.”

This is where a KPI framework can really help.

I don’t want to bore you with too much theory. Essentially, you start at the top and work your way down:

  1. Vision
  2. Strategy
  3. Objectives
  4. Critical success factors
  5. Key performance indicators
  6. Realistic goals
  7. Initiatives

Each level informs the next one, and by keeping it short and focussed, you end up with just a handful of things you rally have to do know, plus measurable (“SMART”!) goals that help you decide whether what you did really helped.

As an example, imagine I have decided that what I really want to do is to run a flea circus. Since I am a wussy, my strategy shall be to start slow, grow organically while still working, then eventually take the plunge when I have learned how much I can charge and whether people would actually want to see fleas performing tricks.

My objectives, in the beginning would be to acquire a flea, teach it a trick, create a reputation, and to take it slowly.

My critical success factors are: fleas actually perform tricks, people appreciate the performance and talk about it, media reports, people are willing to pay at some point, fleas do not die or flee!

The latter is not immediately obvious, but if I am not able to train and use a flea before it dies or runs away, I have no business, right?

Now I can take my 3 objectives, and attach KPIs, in the real sense. They are key, and they tell me whether I’m on the right way or not.

Objective #1 – Train a flea to perform tricks: # of different tricks, life expectancy of avg flea, avg training time

Objective #2 – Create a reputation: # articles in media, good feedback (NPS)

Objective #3 – Grow organically: # of spectators, avg amount people are willing to pay

A simple dashboard containing those KPIs will immediately tell me where I am, and whether I should press on or give up.

KPIs are even more helpful when you have realistic goals for them!

So, a couple months down the line, everything still on track, I might set the following goals for the next three months:

Objective #1 – # of different tricks: 12
Objective #1 – life expectancy of avg flea: 3.5 months
Objective #1 – avg training time: 2 months
Objective #2 – # of articles in media: 2 / month
Objective #2 – Feedback on performance: 80% favourable
Objective #3 – # of spectators: 100
Objective #3 – Avg amount people are willing to pay: CHF 50

Depending on where I am right now, that means I have to do some things, the Key Initiatives.

In my case, those are: teach 2 more tricks, contact another (bigger) news outlet, and schedule first paid performance.

There you go: focus, simplicity, data-driven.

If you work in a company where KPI frameworks or any similar tools are being used, if you know what the overall KPIs are, you can focus on metrics that inform those KPIs. You can provide analysis that is relevant. And you won’t get bogged down by ad-hoc requests.

Look at this example in PDF, if you like: KPI Framework Basic Example Flea Circus

Giza Plateau

With that, we’re back to responsibility, and to layers.

The responsibility for the top three layers (vision, strategy, objectives) lies with the board, c-level, or owner of the company. Everyone else should work towards those, and have their own goals help.

While this is freaking obvious, I am convinced that of those of you reading this right now, 80% couldn’t tell me what their company’s vision, strategy, or objectives are.

Frankly, I would struggle.

That is really lame, even if I hide behind the fact that I am NOT employed by Adobe as an Analyst. I should still know, and it would still help me with other decisions that I make in my daily work.

(Homework: find out! This week!)

Next step: realise that you, too, have a KPI framework for your role.

Your friendly marketer has a KPI framework. It includes things like the pipeline (how much business she is able to line up for sales), brand awareness (how many people have heard of your company), brand appreciation (do people think we’re “good” or not), and others.

It is likely that her objectives for the site are very different from yours, right? I’m guessing hers are more holistic. Where you worry about page load speed and site stability, she worries about how happy people are with the site. One includes the other, and neither helps without the other.

Across the company, there will be a lot of pyramids, all derived from the overall KPI framework.

You (you!) should know what your pyramid looks like, and it should give you a pretty good idea what it is you should be doing.


The fun thing about all of this is that the top pyramid, the company vision, strategy, and objectives, are the “truth” for everything that you do. And for everything that your friendly marketer does.

You might have situations were her and your instinct, wishes, and even objectives are badly aligned, or not at all, but you should always be able to go back up to the top level, and it should help you find a compromise that helps the whole company.

in a recent assignment, I was sitting down with a customer, trying to figure out vision and objectives for individual roles within a new project. We found a lot of aspects where those visions might be opposed, i.e. what one role would want would be bad for another role.

A simplified example: a sales person would like to be able to offer a product at the smallest possible price (so people buy a lot), while the person getting the product would want to spend a decent amount (so they don’t have to bargain that hard). Someone must be responsible for the margin, and they must keep both those roles in check, make sure they work together.

Back to our site, and Marketing handling Analytics, Testing, and othe tools.

Marketing needs the site to perform. It is often their cheapest and most flexible inbound channel. They do sometimes break it, sure, but they need the site more than you (the developer) do.

I guess what I’m trying to say is: trust them a bit, let them do stuff. Don’t try and lock everything down, and certainly don’t handle things that are not aligned with your goals.

Analytics and Target are not successful if you measure site stability, or page load performance. But they can be incredibly successful if you measure other goals.

One thought on “Responsibility – best served in layers

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.