Interview with Danny Mawani Holmgaard
Yulia Tkachova, Product Manager OWOX BI
Olha Diachuk, Creative Writer @ OWOX
Yulia Tkachova: Let’s get acquainted! Tell us about yourself and your experience.
Danny Mawani Holmgaard: I’m a Lead Analyst at the digital marketing agency Impact Extend. I work not only with Google Analytics but also with data engineering, setting up data pipelines, integrating online and offline data, and building data models. I combine and analyze data so it can be utilized both strategically, showing if we’re on the right path or something needs to be changed, and operationally, making sure that the data is in the right systems so it can be acted upon.
YT: How did you start dealing with marketing analytics?
DH: I actually stumbled upon this career path by accident. At first, I was studying digital concept design and wanted to be a web designer. I added Google Analytics to one of my school projects for fun, and suddenly I realized how much you can influence the whole strategic process from the design choices to creating a better experience for users while improving the business results. One thing led to another, and I have worked with data for the last six years.
YT: What do you like most and least about your role?
DH: I get really excited about new projects and designing a solution from scratch. This means going from scoping the necessary data sources to writing the code to connect to different databases and APIs. Other areas I really enjoy are seeing that our efforts are actually making an impact (pun intended) for our clients by giving them the information they need but never thought was possible to have in order to drive their business forward with data-driven decisions.
The least “fun” part of my work happens when something breaks and it needs to be fixed. It can be quite difficult to get data after events have happened that cannot be tracked. Unfortunately, it is never possible to foresee everything that can go wrong when collecting data. This can be anything from a deploy gone wrong to someone configuring a setting incorrectly in their system. The best you can do is to set up monitors for your data and do a spring cleaning once in a while to make sure that you keep up good performance and learn from mistakes.
The meaning of actionable data
YT: What does making data actionable mean to you?
DH: Making data actionable is something that we try to achieve for all our customers. For me, actionable data is information you can take and use actively within the organization, whether it’s something complex such as calculating the customer lifetime value for customers across all channels and then using that in your marketing automation systems, bidding on ads based on the cost margin and sales probability of products, or something as simple as creating audiences in Google Analytics that help you do relevant retargeting. All in all, making data actionable is about making sure that the information you collect can be utilized to dynamically make changes based on actions performed by users in order to drive conversions or provide a better user experience.
YT: What skills and software do you suggest for making data actionable?
DH: It mainly depends on your team’s resources and ambition. The most important thing is to find out what you’re trying to achieve and set goals. You don’t have to work with large datasets or be able to do advanced statistical programming in order to make your data actionable. Start simple and build from there. It can be anything from making sure that your analytics data is measured correctly to uploading refunds or cost data in order to learn which products you’re losing money on.
In general, use the data you currently have available and start small. If you start on too big of projects before the organization is able to handle it, you risk wasting valuable team resources on a project that could fail. In my opinion, the most important soft skill is the ability to be curious and creative in terms of what data can be used to drive the business forward. All the technical actions come second, as they aren’t useful if they can’t be applied strategically.
YT: What about the tools that allow you to act on data?
DH: In order to reduce spending on products and enable our clients to allocate their resources for strategy and execution, we’ve built our ETL [extract, transform, load] processes ourselves using the Google Cloud Platform. This makes it easy for us to scale at low cost. We primarily use R, Python, and SQL for our data processes, but we also use a few vendors to help us out, such as funnel. io and OWOX, as they’re able to handle a lot of data connectors from ad platforms at a low cost directly out of the box.
The disadvantage with building a setup like ours is that it has taken us a long time to build efficient models that can handle large datasets and advanced statistical models that we feel have a good enough quality to productionize. I am a huge fan of ETL tools like Azure Databricks, Alteryx, and similar products that can reduce a lot of the data engineering work with built-in features. These tools, however, can be very expensive and still require a lot of know-how in order to produce results. If you’re going to use an expensive tool, remember to utilize it properly so it can provide revenue instead of draining your budget.
30 handpicked Google Data Studio dashboards for marketersDownload now
Small talk about LinkedIn rumors
YT: A few days ago, we both liked Mikko Pippo’s post on why marketers waste money. Are there any other points you would recommend adding to Mikkos’ checklist? Why do marketers waste their budgets and how can you avoid this?
DH: Mikko is one of the smarter guys I know, and if you are not following him, I recommend you do so. If I could provide an addition to his post it would be to try and enrich the data you already have, such as data on customer loyalty points, refunds, or customer segments. It’s fairly easy to do if you have the client, user, or transaction id present in your analytics setup.
The reason people don’t have the right grounds for data collection or are not using the data they have is as Mikko says: lack of education. If you don’t know what the data can do for you, or if you don’t know how the numbers should look in the organization, it’s very difficult to do anything.
Analytics can be scary and difficult to work with. Most of the time, it’s not that the people sitting with the data don’t want to use it. It’s mainly a question about that person not having time to scale their competencies. It’s therefore up to the organization to build an environment where people with access to data can learn how to use analytics or find an external agency to support the people making the business decisions.
How to become a data-driven marketer
YT: What’s your advice for marketers who would like to become truly data-driven professionals?
DH: Definitely start small. Learning the basics of statistics and being able to do some basic Excel can take you far. Once you’re comfortable working with basic datasets and combining data, you can scale up in the manner that fits you best. When I am learning something new, I always try to match it with a current project and try to learn something new that will improve my work while doing it. I started doing statistical programming with R because I was doing the same job each month that took me several days. By learning something new, I reduced this to one hour. One thing I would like to mention here is that you should not be afraid to exceed the project scope. Even though the hours you spend learning something reduce the total profit of a project, you can still apply everything you learn to the next projects you create, making them better and more stable than if you hadn’t spent the time learning how to do it.
My last advice here is to pick a path. You can’t do just anything if you want to do it well. I am definitely not as good at statistics as our head of business intelligence, and he’s not as good at data engineering as I am, so we have our defined roles but are still able to support each other in the basics of our work. Collaboration is key if you want to succeed, and it’s never a bad thing to hire people from the outside to help scale your department.
Resources & inspiration for analysts
YT: Who is your biggest inspiration in the #measure community? Could you tag a professional whose opinion is worth sharing on our blog?
DH: Mark Edmonson is definitely a guy who inspires me on a daily basis. He gives so much to the community, is always helpful, and keeps making great solutions that help analysts all over the world improve their work when using the Google Cloud platform.
YT: What are your five favorite events for analysts?
DH: Here they are:
YT: Could you name your top five blogs or other sources for analysts?
DH: With pleasure:
… Maybe also my own blog ;) There are so many good resources I also want to mention, such as the #measureslack and Google Tag Manager Facebook group. The best part of the analytics community is that people are eager to share great solutions.
YT: What trends do you observe in analytics and what trends will persist and evolve in the future?
DH: Server-side tracking is definitely something that people will start implementing across organizations, as Intelligent Tracking Prevention (ITP) makes it difficult to collect accurate data otherwise. In terms of ITP, I am a bit ambivalent about this. On the one hand, ITP serves a great purpose of protecting users’ privacy. Personally, I believe that users should always have a choice in terms of what information they wish to share with businesses. On the other hand, this is also preventing some site functionality from working as intended, such as auto-logins and personalization.
Furthermore, I believe that ITP can have the consequences of making companies do workarounds that force people to give more information than they actually did before. Examples could be making users log in before they can proceed to browse a website or refusing access to users who don’t give their consent on the site. In the current state of ITP, I think that it’s not an optimal fit for either users or companies. The world is still in a steep learning curve in terms of following legislation while still being able to provide a good experience. I look forward to seeing how it will play out.
Another thing to start looking out for is the App+Web analytics platform. It’s still in the very early stages, but it will be something that I highly recommend implementing and building on as it evolves. That way it will be possible to use the capabilities of the platform once you need it. It also comes with a lot of cool features such as free BigQuery export.
YT: What new challenges do you face at your day-to-day work compared to at the start of your career as an analyst?
DH: My role has changed a lot, from being more of a supportive role in the beginning of my career to driving projects end to end. Before, my challenges were to understand how things worked and do my best in order to ensure that the reports were showing the right metrics and that tracking was set up correctly.
Today, I know how to do these things, and the difficulties are a lot about communicating strategy so clients understand the benefits of collecting data and push for their organizations to use that information actively. In brief, I have gone from executing implementation and reporting tasks that were assigned to me to driving the strategy and the “why” for how data should flow. I believe that it’s key to have done a lot of grunt work before being able to understand and drive strategy.
Before you can progress as an analyst and develop your skills, you must master the basics and understand how everything is connected.
YT: Google Analytics or Adobe?
DH: I am personally more into using Google Analytics, as it’s more available to everyone and has a great API. I do, however, think that Adobe has a strong and stable product, with some really cool ways to analyze data, while also being a stronger product in terms of compliance.
YT: Data Studio, Tableau, or Power BI?
DH: I like using R’s ShinyDashboards, as you can build anything you want with any types of graphs you want in a dynamic manner. Also, it doesn’t require expensive licenses (unless you want to use their hosted solution).
To be honest, I think there’s something great with all the tools mentioned above. Data Studio is free and fast, PowerBI can be cheap and provides a lot of great visualizations out of the box, and Tableau has some really cool sharing functionalities and features that are great if there are many people exploring and analyzing data within the organization.
YT: AppsFlyer or FireBase?
DH: I don’t have a strict opinion here. However, the development of APP+Web definitely makes FireBase interesting right now.
To sum up
The key to success for an analyst is patience and a preparedness to start from scratch. Thanks to Danny for this simple advice that helps us stay focused and persuasive in our analytics.
If you liked this interview or have questions for Danny, leave a comment below! Don’t forget to subscribe to our blog to get even more useful talks and articles from Julius Fedorovicius, Brad Geddes, and many more from the analytics world.