Data & Processing | How do I customize my OneView?

Now that you have a better understanding of ChannelMix's data model and how our Solutions are built, here are some examples of frequently asked questions around customizing a OneView as well as our take on whether they are a good practice, a bad practice, or a practice that can be done but can get ugly if you're not careful.

Question: "Can we have Google Analytics (GA) data in a Media OneView?"

Answer: "Not a great idea."

• It’s website data not media data.  You should check out the Web Analytics Solution. This will give you a clean, comprehensive look at your web traffic and even tells you the channels, sources, and campaigns that users navigated to become a visitor. 
• GA data never meshes with Media data
• GA transforms data so that it often doesn't match the media source and thus, will not aggregate. 
• GA data is permanent and cannot be changed. Your media data can be changed. So if you change something in the media source, it will no longer match GA.

Question: "Can I have One Solution combined with another"

Answer: "Not really"

• Each solution is designed to answer specific questions and to do so as intuitively as possible. As soon as you start mixing and matching them, you lose that intuitive nature. Then, instead of showing off how well your campaigns are doing, you're answering questions about why Facebook data is showing up in your Paid Search dashboard.
• Scalability also comes into question on this one. Oftentimes, analysts want to combine two Solutions because they have a user or client that wants to see how two marketing asset classes intersect.
• Example: Let's say you are an advertising agency with several clients. One client is an ecommerce site. They want to see how paid media intersects with their website traffic. If you jump through all of the hoops to cobble together a halfway usable dashboard that shows paid media and web analytics in one place. Only, now their dashboard no longer matches the structure of the rest of your clients' and if something breaks or needs updating, someone is going to have to have the tribal knowledge of how that particular dashboard works in order to fix it. If you stick to (or even just close to) our Solutions, you will have the same repeatable dashboard for all of your clients, making any updates seamless and it is easy to understand.

Question: "I need additional granularity. Can we add 'creative' to my OneView"

Answer: "Sort of! We can certainly help you get what you need here but there are considerations to be aware of."

• Additional data can add latency in reporting.  It can slow your extracts and it can also slow your actual reports, because your BI tool still has to render more data.  In most cases, your BI tool is storing the data on a server and that server can get bogged down.
• If you're close to your storage quota, this could be a factor to consider, since you’ll be adding more data. You may have to buy more storage.
• If your extracts start to time out or they don't run fast enough for your needs, then additional data may force you into needing to purchase more “horsepower” for your database to be able to churn the extra amount of data.
• This can sometimes make the data more difficult to understand. More data points means more to read, comprehend, and derive insights from. It usually also requires more functionality in your visualization as well. For an analyst, this might be fine, but for an executive team, you don’t want them spending too much time slicing and dicing data in a dashboard. You likely want them to be in and out quickly finding the answer to their questions with ease.
• It should be noted that you can have a OneView dataset that is at a different level of granularity than what the Solution uses, BUT it will be a new, separate dataset. The idea is to create the dataset in a vacuum so that the existing Solution doesn’t break.

Question: "Can I please add these new fields that only apply to one data source?"

Answer; "We can, however it is not ideal."

• We’ve gone down this path many times and it’s never great.
• Rows of data associated with a source that doesn’t provide that field will always be NULL. This gets confusing to many people that are not in the data every day. So, when you’re presenting your marketing performance to your Board of Directors, you will inevitably get a lot of questions (not the good kind of questions) that you’ll have to take the time answer and it will detract from your presentation.
• There is also the fact that you’re still able to blend or JOIN the underlying dataset into your dashboard in order to pull in source specific fields if you need them.
• All of that said, we've had to do it before. We actually have this in our Social Post OneView because we wanted to be able to provide the breakout of Facebook reactions aside from just likes. None of the other platforms had those.
• One note on this, we did make sure to name these metrics 'Facebook Reactions HaHa', 'Facebook Reactions Wow', etc... just to be sure that if someone saw a bunch of NULL values for other sources, it would be very evident that these metrics are Facebook specific. 

Question: "I have a special data source that doesn't belong directly to any Solution, but I still want it added."

Answer: "It depends. Send your request to our Data Engineering team and they will help design a data strategy that works with your Solution if it is possible."

• This is similar to the second question above, but gets a little more specific.
• For instance, you may not want to try to add data about your webinar platform within your Paid Media dashboard. This one seems tricky, but if you think about it, your webinar, in all likelihood, is not “Paid”. You likely pay to use it, but the cost of your webinar platform is very different from the cost structure of your Paid Media.
• You may end up comparing apples to oranges if you aren’t careful. To reference my previous webinar example, if your webinar gets 1000 impressions and has no clicks, cost, campaigns, etc…. AND all of those impressions happen at one time (during the webinar), then you might create some pretty confusing reporting. Think about the anomalies this would cause in a click-through rate calculation, for instance.
• Now, all of that said, some webinar platforms provide campaigns, engagements, and other fields that could make this much more useful in a Paid Media OneView. This is just one reason a Data Engineer should review the data with you and provide insight on whether your data source belongs in your OneView or not.
• In some cases, the answer may be "No, this can't go in a OneView." That doesn't mean that you can't still tell the story. Often when this comes up, the solution is to add a new tab to the dashboard, one that only has the oddball data source. It's still part of the story, but should be viewed separately.