Teem Insights is a way for key stakeholders at Teemo to share their feedback and insights with the rest of the world. 


The 3 Levels of Data Analysis

Richard

Okay. So for Teem Insights, I’m here with Raphael Cattan. Raphael is a data analyst here at Teemo. Can you tell me a little bit about what you do in your day-to-day?

 

Raphael

I would say that my job is multi-leveled in the way that I handle and analyze data on different levels. The first level would be in prevision of the campaign. So that’s before the actual start of any campaign. The second level would be during a campaign, in order to make sure that things are running smoothly and that everything has been imported. And the third would be post-campaign. So after the campaign has ended we want to have some sort of complete analysis. That is we want to find some insights into what happened: what we did well, what we could do better, and really, what this campaign brought to our clients.

So I think it’s really on those three levels. As far as, you know, day-to-day events, day-to-day tasks, I’d say that it really depends on the needs of the company and the needs of the clients. You know, it can range from a simple query and data analysis to a full book report. It really is dependent on the needs of our clients.

 

Richard

Gotcha. Okay. And you say you’re doing a lot of work during and before the campaign. When that’s happening and there’s a deadline on a project, what do you find is something commonly difficult during that process?

 

Raphael

Well, you know as with anything technical, you sometimes have some limitations in the computer and in the data source. For example, sometimes you have to wait for data to compile. We have to wait for data to refresh. And when you’re working on a time constraint, this can be a bit limiting. You have to wait for fresh aggregate data to come your way or to be able to analyze it.

 

Coding for Humans & Computers

Richard

Okay. I want to dive a little deeper into the technical side of things and learn a little bit about the types of systems, platforms, and coding languages that you’re using.

 

Raphael

Yeah. so there are a couple of different platforms that we’re using as far as data analysis is concerned. The first one is called Mode Analytics. And this is very easy to use, very visual and allows anyone to really run a report based on something that we have created before.

And so the idea is really to use this platform to have some preset reports that are ready to be launched. So that way we have those reports and we can tailor them to the client or need very quickly. We have these reports ready and anyone can just run them, both technical and non-technical people. It really doesn’t matter. It’s just a very small learning curve and these reports can be very useful. The problem with this platform is that it can have some limitations. 

And so that’s when we switched to the other platform, which is called BigQuery from Google Cloud Platform. And so this is something much more technical and much more in depth. It requires you to create every single query by hand. And so that’s what we use when we need to dig deeper into a report or get some information that is not part of the regular set of analyses that we have.

And so BigQuery requires SQL, which is one of two languages for databases – one of two famous languages for databases, that is. The other one being Python. And SQL is sort of the simpler one. But it’s not simpler in the way you use it, and in terms of technical difficulty. It’s simpler in the type of data it manages and the type of data analysis it runs. The main difference with SQL is that SQL is going to allow you to run analyses that are more natural and more understandable. When Python is going to let you analyses that are much more technical. These Python analyses won’t really be understandable for a human being. Rather, the results of those analyses are going to be used by other computers, that are then are going to use SQL in order to get the information so humans can understand it.

So it’s really the last step in translating from computer language into human language.

 

What’s the Story with Data?

I think that data is extremely important because data tells a story.

 

Richard

So you talked about multilevel parts to your day-to-day. How important is data to the sales process before, during and even after a campaign?

 

Raphael

I think it’s extremely important. If you asked me, I would tell you that data is the source of everything, and the future to every single job in this country. I think that data is extremely important because data tells a story. In the case of before a campaign, it tells the story of what we’re capable of doing and what the future could look like. And during the campaign, it tells the story of how things are going and gives you a weather report of how the campaign is doing. Abd after the campaign, well it’s a story in the more basic sense. Data tells the story of how the campaign went. It tells you exactly what happened. You can understand if a campaign was successful or not, depending on the metrics you use. 

Data really allows you to understand what happened. Because if you think back to a few decades ago, there was no way for anyone to verify if a campaign was successful or not. There was no way for someone to know if a TV spot, for example, actually drove visits to your store. It was all empirical. You didn’t know. You didn’t actually have proof that it was that TV spot that drove business to your store. And that all changes with data. And so I think that data really gives a very new and important edge that has never been available before.

 

The Source and How You Use It

Richard

I think that’s a really good point. The proof is in the data on that. On that line of thinking, one thing I’ve heard (from salespeople and developers and our CEO) that differentiates Teemo is that our data is the best. It’s the most accurate. So what do you think it is about our data that differentiates us?

 

Raphael

So, you know, I specialize in data and I’ve always said that there are two things that are extremely important to good data. These are: the source of the data and what you do with it. And as far as number one is concerned, I think the source of data is the most important part. Your data is only as good as the source. That is a real issue when you’re handling your data. Because if the source of your data is bad from the beginning, then you will have bad results, and you will have a bad understanding, and nothing will be accurate. 

 

Product-Market Fit from a Data Analyst’s Perspective

Richard

And just one more question. I know that you recently completed sizing analyses for various segments. Having done that and looking at all these different verticals, what do you think are a couple underrated verticals that could really benefit from location-based advertising?

 

Raphael

I think that’s an extremely good question. But I also think that it’s a very constraining question because I think there are a lot of verticals that could actually benefit our service. But if I had to choose two, I would say that the food industry – what they call QSR (quick service restaurants) – could tremendously benefit from our technology, for one. Also, retail in general has a huge opportunity. And, the bigger the retail store, the more potential there is from working with us.

So for example, a company like Walmart would really benefit from our technology because we would be able to provide insights that they never had and probably never will. And because of their size and because of the importance of their brand and their retail network in the US, that is the type of company that can gain the most from our offering. 

 

Richard

Awesome. Yeah, I agree. So I really want to thank you for joining me today and offering your insights. I’ve learned a lot.


Raphael, the Data Analysis Pro

Raphael is a Data Analyst at Teemo with many years of experience in data technology management and development. He has a knack for photography and a deep interest in aviation. You can find him on LinkedIn.