Analytics come in many different forms – business intelligence, descriptive analytics, optimization, predictive analytics, forecasting and more, even at the highest level – but deciding what types of solutions you need in a particular situation is the work consultants do all the time!

The key to creating an effective prescription is to listen to your client. Why are they upset? What are their key sales metrics? What are their most immediate plans? What long-term goals do they have? And how could we reach their goals without using too many resources and without causing any further problems? Building a sound roadmap for data and analytics is very important, because no one wants to look back a year later with the realization that they could have done something different. Reconstructing any line of a business is expensive, but it happens all the time. Many marketers and decision-makers jump from one idea to another because they are desperate and under a lot of pressure. It is unfortunate when users realize they do not know how to swim until after they jump into the pool.

But why does this happen all of the time? It happens because there are far too many patients. When it comes to analytics and data, many marketers are not experts. They have a mid-level understanding, but they are not experts. Sometimes they need to call for assistance when they reach a certain point in the process. But the problem is that bad patients are often the ones who come by with self-prescribed solutions. There was a time when I spoke with a client who asked for a neural-net model as soon as we started talking. I told them to think about their decision a little bit before making any demands. Ask the client to go through a discussion about their business case before they spend a ton of money on a strategy that may not work.

There are opposite cases too. Sometimes marketers are so insecure about their data assets, or how they understand these assets, that they are not willing to hear about solutions that are a little bit complicated or difficult. Even if they really need these solutions, they simply want something simple they can implement. They say something along the lines of: “We have incredibly messy data sets and we cannot possibly go through something so statistical.” It sounds like a patient who needs some treatment in the ER but is too scared to go through the process. Doctors need to be ready, not the patient.

Messy datasets are not an issue at all. If we wait for the perfect set of data, we would be waiting forever. Sometimes we need the help of specialists for the specific purpose of going through complex and messy data in order to find a solution. They use statistical techniques to figure out the best approach in each situation.

Analytics is all about looking at a situation and making the best of it. Cleaning up messy data is the job and it is never an excuse to not move things forward. If people think that simple reports did not require data analysis and cleansing, because the end result looks so neat, they are completely mistaken. Most datasets are not plugged into an analytical engine right away, because they require cleaning up and sorting through beforehand.

Besides, you need so many different types of analytics in order to get through the various business challenges, and analytics are not meant to happen in a preset order. We can get into predictive modeling because this is what the business needs, not because a marketer just took a Reporting 101 class and decided they are ready to move onto Analytics 202. I often talk about the fact that deriving insights from simple reports is sometimes a lot more difficult than creating models or managing complex data. On the flipside, marketers should not get into advanced analytics simply because they are curious. Every analytics activity should be justified with a business purpose. Follow this up based on the strategic data roadmap, not based on the difficulty of each task.

The main reason why it is so difficult to capture the entire essence of “the right thing to do” when it comes to data is because there are so many viewpoints to consider at the same time. Now we can take a look at some essential dimensions:

  • Industry: Sometimes consulting companies and service providers center their practices around industry verticals. I do not agree with the approach they take, since there are many more critical dimensions I am going to list in the article. Nevertheless, the industry that a company is a part of – whether it is finance, banking, entertainment, retail or publishing – plays a very important role. Not only are the business models different for various industries, but the shape of collected data and the success metrics mean you need to use different types of analytics. Aside from the industry breakdown, we must also focus on B-to-C and B-to-B cases individually.
  • Marketing and Sales Cycle: Even when we have companies in similar industries, we can observe some very different marketing practices among those companies. Some marketing organizations are sales oriented, and they require different solutions as opposed to companies that are more marketing oriented.
  • Marketing Channels Employed: I have been talking about the relevance of buyer-centric views with regards to marketing, but the truth is that most companies break down their marketing departments and activities based completely on channels. For this reason, channel usage is a very vital factor to consider, since each channel results in a different type of data and needs different messaging techniques.
  • Target Buyers Lifestyle: Buyers are going through various cycles of being a consumer when they interact with a company, and each stage results in different data. For instance, the available data does not even look the same for prospecting and CRM stages when we analyze it from the marketer’s POV. For win-back programs, marketers need to sort through aged and third-party data.
  • Data Availability and Shape: As I said earlier, some data sets are going to be messier than others. Marketers should never give up because the data is not sorted through easily, and disorganized datasets are nothing you need to run away from. And the truth is that organizing data sets and consolidating disparate data sources can really help to create a 360-degree view of how things are going.
  • Existing Teams and Divisions: Various divisions have their own agendas, which mean it is sometimes tough to navigate the political landscape between divisions. But this is part of the data player’s job and they must ensure these varying agendas do not become bottlenecks.
  • Level of Sophistication for the Users: Users who are intermediates, experts or novices are going to need a different type of solution. It is not because they have to go through certain analytics courses in order, but because marketers have to make decisions based on their relevant skillsets. Some marketers may want to get into an analysis of sophisticated data, while others may not even want to go through more than a few pages of a data report.
  • Pain Points: Each and every company has pain points, even the most advanced ones. The important thing is to fix problems without giving up of the long-term goals. Far too many analytical solutions end up being near-sighted, and far too many projects are created to deliver quick results. This leads to “oh no” moments where everyone realizes they let go of the big picture in order to fix a short-term problem. All data projects must have a clear roadmap, even if the goal is to solve a small problem.

Analytics consultants have many factors they need to consider when they are coming up with solutions. But, from the user’s POV, data is something that needs to be accessible easily, even if the professional data players went through hardships and complications to sort through the data. In the same way that we take our daily weather reports for granted, I am sure we can accept that understanding and analyzing the weather is not an easy job. Data scientists need to understand that the users are going to demand the following things from data:

  • Easy to understand and relatable for everyone, not only the experts.
  • Small answers in bite-sized formats, not tons of information that is unrefined.
  • Reliable, accurate and effective on a consistent basis.
  • Available almost all the time, not on an infrequent basis.
  • Easy accessible through the channels and devices users want to use.

When I talk about “Smart Data,” this is what I am talking about. There are so many types of users, business models and challenges, along with plenty of clean and messy data. Navigating through all of these hurdles is the job of the data players. Good data scientists do not complain about users, even though they sometimes need to tell them they are headed down a wrong path. At the end of the day, even the worst patients deserve the best treatment!

About The Author

Leave a Reply

Your email address will not be published.