Hello blog-o-sphere. I am Amy Sharma the Executive Director of R&D here at Social123. I’ve been asked to write a blog about “what is DaaS?” The funny thing is that right now DaaS is evolving from a buzz word into a thing. Thus, there is no real standard definition of DaaS. Here is my interpretation of things. I realize I am so influential, that this will go viral, and from here on out be the standard definition of DaaS. [See Infographic Below]
To set things up: in general, like most marketers, you are sitting on this pile of data. And your real job, the one you want to do, is to get the wonderful creatives you developed out to the right audience. You really don’t want to mess with managing and cleaning your database – that’s scary territory. You just want to create some segments and communicate your messaging to the proper audience.
Someone told you that you could use data analytics to do this, but what is data analytics anyway? I am going to define analytics, and then I’ll define how you do it, and then I’ll talk about Data as a Service.
Now, some definitions. Stolen from Wikipedia.
- Analytics is the discovery, interpretation, and communication of meaningful patterns in data.
- Business Intelligence uses descriptive statisticswith data with high information density to measure things, detect trends, etc.
- Big data uses inductive statistics… to infer laws (regressions, nonlinear relationships, and causal effects) from large sets of data …to reveal relationships and dependencies, or to perform predictions of outcomes and behaviors.
- Descriptive statistics summarize the sample. Inductive statistics learn about the population the sample represents.
In the great wide world of data analytics – there is a hierarchy of things you can do with your data. They are listed here from least to greatest.
I am sure that you noticed that nowhere in any of these definitions are the words “take action.” And that’s the rub. You need to do something with your data. You need to take actions. You don’t care if the analysis was descriptive or inferential or causal. You just care that the results of the analytics enable you to do something.
Alas, if you hire a data scientist, given the above definitions, they can analyze your data, and leave you with this:
Well that isn’t what you wanted. The first step to foiling their nefarious plans is to understand the analysis process. Here it is in one simple chart:
Looks easy right?
Here is how that goes.
- Decide what you want to do. Run a campaign? Build a segment? What?
- Inventory the data you have and need to answer the question.
- Clean / augment / process / de-dupe / normalize / merge this data. During this process you will learn that most of your data is very messy, and lots of it is either junk or unnecessary. Without clean data everything else in your process is moot. It becomes a garbage in / garbage out situation if your data isn’t clean or valid.
- Ignore over 90% of your data. Cry just a little about that and then move on.
- Using rules and algorithms, perform real-time analytics on incoming data. This leads to real time actions. Examples:
- Based on browser history decide instantly what type of ad to plane and where
- A/B email campaigns
- After your immediate actions, store the data and the results of the actions
- Do some offline processing of the data. During this processing, you develop the rules / machine learning / artificial intelligence that are then used in real-time analytics (step 4).
- Debrief and analyze your results. Did you get what you wanted?
- From this you can ask new questions so to refine your needs or execute different goals.
That’s it. Simple no?
Okay, so some of you, by now, don’t want to do data analytics. You don’t even want to do data management. You just want to hand someone your data, have them clean it up, write some rules, and then bam! you just use your data to take some actions.
So what can you offload?
Let’s look at this picture again. Now with a descriptive green box.
Everything under that green box can be managed elsewhere. I left “what do you have” as a half-portion, because you should pay attention to what you have, but someone else can manage it.
The green box, my friends, is Data as a Service. It’s all the management, and processing, and rules, etc. that you don’t want to do and don’t need to do. You just ask your question, see what data you might need to analyze it, and then use the results. Everything else is left to the professionals.
So what are you responsible for? Deciding what you want to do. That’s very important. We all have job descriptions and expectations. And the DaaS can do anything, but it needs you to provide the mission.
You also must analyze the results. This keeps your DaaS on track. And makes sure that you aren’t using pirates to combat global warming. We saw earlier that data analysis doesn’t mean take action. So looking at the results of the analysis results in your taking action. That action could be to run a marketing campaign with a well segmented list. Or it could be to fire your DaaS because it stinks.
Your call. You are in the driver’s seat. You ask the questions. DaaS provides the answers so you can take action.
Check out our DaaS Infographic, too: