A Manifesto for Analytics and Optimisation Consultants

This is my view on how SaaS consultants should work, formulated as a bunch of principles in the style of the Agile Manifesto.

I like the style, because it recognises that all work happens on a spectrum, and that good work can come from knowing the broad direction, or which way to go when in doubt. The “a over b” style feels to me like it covers all the complexity, yet still tells me what to do if I want to do it right.

These are for Analytics and optimsation folk. And this one is more about the technicalities, while the “Value Manifesto” in Consultants (can) do Everything is more about general work ethics and guidance.

So:

Standards over customization

Keep it simple, cover 80%, only customize when absolutely necessary.

Our solutions are complex, and when combined, they are so complex that even we struggle with it. Try to keep everything you do simple, make it easy for users to understand what they do, and how to do it. Only ever customize if it makes something easier.

Restraint over Quantity

Only track what you actually need. Check with your KPI framework.

Doing too many things at a time tends to overwhelm people. Limit what you do to only those things that actually make sense and create value for the customer. It makes sense to look from the top, use a KPI framework or something similar.

Quick over Complete

Get results quickly, rather than going for a “big bang” approach.

Try to show value as quickly as possible, rather than trying to sort everything out in one fell swoop. A customer who sees value is more likely to work well with you one the rest of the bucket list, and beyond.

Testing over planning

Don’t follow instinct, instead test what works.

Use testing when and where you can. Assumptions can be wrong, instinct can be wrong, even your taste can be wrong.

Insights over Reports

Don’t implement so you can report. Dig for insights in what you have.

It is better to find insight in existing data, no matter how dodgy that data is, than to implement something new in the hope that it will be clean, useful data.

Action over Insight

Better to have “form abandoners” segment and target that, than to have form abandonment analysis.

It can sometimes be easier to segment, test, and then target, than to analyze and understand. We have a tendency to want to understand, but consider this: testing or targeting on web sites can be quick and insightful. Go for action rather than insight.

“Why?” over “Let me think how we can do that”

Don’t just do what customers say.

Even if it means that you have to eventually back down, you should still always ask “why?” when a customer asks you to do something. Think about the bigger picture, the strategy, don’t be a reporting worker, or an implementation bee.

Guidance over “What do you want me to do?”

Better to guide customers towards a solution that works, than to do exactly what they say.

Customers can be caught in their tactical world of specific tasks and things. It should be our aim to get them to see the big picture, to tell them what makes sense.

WE are the experts.

Spell it out over buzz words

Always define terms, or avoid them completely and describe what you mean.

Our industry is young, our solutions are complex. If you spell things out without using buzz words, customers are much more likely to understand you. And real dialogue leads to better work.

Notes

The above are based on what I see colleagues, partners, customers, and of course myself do, and what I think is harmful to both us, and our customers.

Our customers trust us. If you are like me, this is a little shocking every time it happens, but it is the truth, and we do merit it.

I feel that with the above, we can be sure to get to the best result for them, and for us.

11 thoughts on “A Manifesto for Analytics and Optimisation Consultants

  1. Lots of great points Jan. My top one has to be Standards over Customization. I need to maintain things over the long haul, and often rely on people with varying levels of technical expertise. So while it may seem smart and efficient to make some fancy custom stuff, in the long run that supposed efficiency usually leads to confusion and problems.

    Like

  2. Great list!! Especially the first 2 points are usually the root cause of most frustrating situations.

    Not sure I agree with “Insights over Reports” point, as I believe optimizing data collection for reporting requirements make sense from the end-user perspective. Maybe the key is the right balance.

    Like

    1. The right balance is magical, agreed. I have two reasons why I think this holds, interested to know your take on those:

      1. a report in itself is worthless. The simple assumption that stuff will stay the same and the report will be useful more than once goes against the drive to make things better.
      2. a little bit of insight based on “rough data” can drive more change than a super-clean report. I see a lot more value in exploring, diving in, thinking, understanding data, as opposed to setting up things. At some point, setting up has to stop. IT has no value in itself.

      What do you think?

      Liked by 1 person

      1. 1. I’m with you on that, but there are a lot of cases (usually with higher management) where you need to have a report/dashboard. At least for a few basic metrics & dimensions.
        In these cases, I’ve seen that it helps reduce frustration if you can “beautify” collected data (meaning, implement so you can report) to make reporting easier for analysts who are not familiar with technical difficulties or make reporting inside Adobe analytics easier. A few examples of collected data beautification could be: fixing common typos in collected values, fixing the format for values coming from different website platforms, renaming data layer values assigned by developers to something that makes sense for the business.

        2. I also agree, but again when you are addressing users who just want a dashboard in workspace, sometimes you need to just simplify “rough data” to allow efficient reporting, without the need to extract data out of Adobe analytics and into a tool allowing you to work through “rough data”.
        An example I can think here is the browser’s user agent. Having the complete user agent in a dimension can be very helpful to extract insights (e.g. detect bots), but a business user will probably get overwhelmed by seeing the complete user agent string and would just prefer to have a dimension showing the browser or the device type.

        Again for all of the cases above, it brings a lot of value to know the end user and to understand the “Why?” (As you already said 🙂 ), to find the right balance.
        So I wouldn’t say that I do not agree with your “Insights over Reports” point, but I wouldn’t consider it as a strict rule that can apply to everyone.

        Like

  3. Great post and I enjoyed reading it.

    Though I like all of the points. I would like to call out the Quick over Complete. I was recently reading Kaizen Methodology for lean practice that says a “quick” good solution is valued over the best one. Realizing the value quickly can gain huge trust from customers. I hope you are trying to refer something like phased implementation rather than a “one go” approach.

    Like

  4. These guidelines make a lot of sense in my experience, Jan. I love that you’ve simplified what people can debate over for hours into a few statements. I am a fan of simplifying, and see that you reinforce it several times in your comments: “Keep it simple, cover 80%, only customize when absolutely necessary”; “Spell it out over buzz words”; and “Only track what you actually need.” I also like your focus on insights and action rather than reporting and analyzing for the sake of it.
    I am especially interested in your post because I’ve just written my first blog about best practices in displaying analytics to encourage action & decisions instead of endless analysis (https://lizzieflippit.com/analytics-that-drive-insights-and-action/). I call for what you call “restraint over quantity.” For example: “Remember that your display is not a Christmas tree, and resist the urge to add pointless trimmings. Less is more. White space is good.” It would be great to know what you think of my recommendations if you get the chance.

    Like

    1. Hi Lizzie,

      I finally got around to reading your article. Very good points!

      This one I liked a lot: “Each view doesn’t have to address every single need, but every need should be addressed somewhere.”

      To me, that works towards simplicity, and that is always a good thing!

      Thanks for sharing!
      Jan

      Like

Leave a Reply

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

WordPress.com Logo

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

Google photo

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

Twitter picture

You are commenting using your Twitter 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.