Introducing Your New Competitive Marketing Advantage: Duco Analytics
January 3, 2021
Let’s cut to the chase. We have a new product to announce.
It’s a first for Duco and we've been waiting a while to let you know.
Anyways, without further ado, I’m ecstatic to introduce……..*drum roll*.......
For those day-one Duco fans out there, you’ll know that we started off as a customer experience management firm. Our goal was to help clients provide the best possible experience to their customers. When you produce results, clients like you. And when they like you, they want to know what else you can do.
Our foray began with paid social media ads. Facebook, Google, and LinkedIn to be specific. Our efforts paid off, and clients began to see tangible results. More web traffic, leads converting, increase in sales, the whole nine yards. Social Media advertising was (and still is in my opinion) the most affordable advertising channel out there. But as our client base and data resources grew, our control diminished.
We started to get overwhelmed by the data coming our way and needed a process to crunch these data in a presentable and meaningful way.
You see, clients like you care about the bottom line. Discussion of our latest successes with geo roll-outs and retargeting put them to sleep, save a few, but when we shared direct qualitative data relating to their KPIs, ROIs, and overall growth, their eyes lit up.
As much as we espoused the philosophy of experiential marketing, the numbers need to add up. By aggregating the ocean of data we’d receive on a daily basis into a cohesive, visually-pleasing format, we could strongly demonstrate our value and keep clients happy and confident. But of course, we also realized this was a tremendously difficult task to maintain manually. We needed a dream team to help us build the perfect data-crunching, number-analysing, insight generating machine. This would require a wide array of skills from data science, data visualization, marketing expertise, coding APIs, and (much) more.
Attempt #1: Excel
A little intimidated by the “coding APIs” part from before, we turned to our trusty old spreadsheet maker: Excel We would spend hours and hours preparing reports and realize quite frequently that one small error would throw everything off. Visually, it wasn't too pleasing either. Combining a myriad of cross platform data was difficult enough, and making it look good was damn near impossible. So, we re-assessed our efforts and decided to find a better solution. Preferably one that we could
download in time for our next client meeting (no pressure).
Attempt #2: Existing Software
Our journey started, as always, by googling solutions. We examined and reviewed every off-the-shelf marketing analytics software we could find. Unfortunately, there were few that met our needs and the ones that did couldn’t connect to all the data points and present them in the way we wanted. They were all SaaS models – which is probably a great thing for the VCs backing those companies (yes, I said it) but we didn’t feel they would provide the type of customization that we needed to make our client reports effective. And so, in 2019, we decided on our third and final solution: create one from scratch.
Attempt #3: Duco Analytics
If we were really going to make this happen, we couldn’t leave any stones unturned. We spent countless hours talking to experts from all fields related to data science and computer programming. All of this research and development led to one grand ambition – how can we automate and generate insight from marketing data using machine learning and artificial intelligence?
The mass of data available to marketers and the importance of the problem we were solving (generating business for our clients) seemed like the perfect combination of variables for an advanced analytics solution. Of course, this was not a new problem for us. We wanted to know what it would take to develop an analytics platform – whether it was possible, and how much it would cost to get going. We spent about 9 months researching and developing the
**Cut to a few months later** We realized this was a feasible project and took another bold move: we sold it to one of our existing clients before it was finished (what can I say, we’re good marketers). Thanks to our Scientific Advisor, Colin Kemp, who always brings us back to the basics of analysis (data visualization, correlations, and frequencies), we decided to initiate product development to solve two basic challenges:
1. Can we connect to all the different platforms our clients are using and store it in one database?
2. Can we do it in a meaningful way?
A third challenge we were facing personally was finding the right coder for the job. As a 4 person
marketing team that could barely run a Zoom call, we knew we had to hire someone with technical
Okay, small side story here from one of our founders:
I’m laying in bed one night thinking I need to hire an expensive programmer, need to develop a job description, interview and test dozens of people…if you're a startup founder or entrepreneur who also has a full time day job, you know that this is not an easy task. I kid you not though, as I lie there thinking I get a LinkedIn notification – Bingkai Liu would like to add you to his network. I open the profile and there is a photo of a young grad from the same university where I studied my undergrad (shout out to University of Windsor). What does the summary and experience say? Graduate level expertise in python and data science seeking employment in Ottawa. Sent him a message that night at 11:30, met with him the next day, and hired him the next week. Kid you not this happened.
OK, back to the story:
Once Bingkai (our resident python developer and data scientist) solved challenge #1 (collecting data from all the platforms our client data was connected to) we were able to initiate the data visualization process. Challenge #2 was complex and we certainly spent more time trying to perfect the data analysis and visualization. This is where Colin (our scientific advisor) started guiding the process. And believe me when I say this, the results were remarkable.
Let me give you two examples that blew our minds:
When we connected to the Google Ads platform and started visualizing the conversion data, we realized that one of our campaigns (a Google Smart Campaign) which we had not paid too much attention to recently, was blowing up with conversions. In fact, it had converted over $40k worth of sales in the past 3 months for our client. Google’s dashboard was just not intuitive enough to show the conversions in a meaningful way, mainly because the campaign was a smart campaign. Also, we were hyper focused on two separate campaigns which we had initiated recently and we thought would be the highest performers. But there was no hiding from the visual – it was crystal clear that we had to focus our attention on this ad and so we did. With further optimization of the keywords and ad spend, we were able to increase our ROI significantly.
Let me give one more example that happened just recently. I opened the weekly dashboard and our total ROAS (return on ad spend) had dropped from 4.28 to 1.98. Now, the first amazing thing is that we even know what our total ROAS is – we were able to know this because we combined all of the differentad platform data into one database. The second amazing thing here was that we were able to see the sudden drop and look into it right away. The significant drop was related to the same ad campaign that was blowing up only weeks before. We dove deeper and figured out what was going on – nobody was searching for this product in our target locations – probably because it was the middle of summer and everyone was on vacation – so what did we do? We expanded our search locations to include broader geographies. Once again – Duco Analytics gave us the insight we need to take action.
To summarize, our vision is this: Get to that point where Duco Analytics will not only find and predict issues before they arise, but react independently to solve the issue. That client we first sold to? They use Duco Analytics religiously now. Every weekly meeting is 5 mins of small talk followed by 55 mins of observing charts, comparing data, and discussing potential correlations. Clients go nuts for this stuff. If you’ve made it this far, you’re probably in the same position we were 2 years ago. You know what to do.