In 1990, the internet took on its most recognizable form. It brought connection, knowledge and speed that was previously inaccessible. Fast-forward 27 years, and I get asked a lot about the most recent form of the internet – the internet of things (IoT). And while I think the current possibilities [...]
In a recent presentation, Jill Dyche, VP of SAS Best Practices gave two great quotes: "Map strategy to data" and "strategy drives analytics drives data." In other words, don't wait for your data to be perfect before you invest in analytics. Don't get me wrong -- I fully understand and [...]
The insurance industry is becoming increasingly focused on the digitalization of its business processes. There are many factors driving digitalization, but it’s clear that a reliable and meaningful database is the basic prerequisite successful digitalization strategy. Insurance companies are increasingly prioritizing digitalization, not because this issue is currently "in," but […]
While men still outnumber women in the analytics field, there are plenty of opportunities available for women. At a recent Chief Data and Analytics forum, I was encouraged to see a well-balanced number of senior executives presenting about the business of analytics. Speakers included 12 women and 14 men, which indicates a […]
Recently, I was talking to a director of analytics from a large telecommunications company, and I asked her, “Do you think we have a skills shortage?” She replied, “NO, I think we’re just looking in the wrong place.” I wanted to hear more as this analytics expert may have just […]
As support analysts in the SAS Technical Support division, we answer many phone calls from SAS customers. As members of the SAS Foundation team, we get questions that vary significantly in content from all of the areas that we support. We offer coding tips and suggestions as well as point customers to SAS Notes and documentation. A common question that we are frequently asked is how to collapse a data set with many observations for the same BY variable value into a data set that has one observation per BY variable value. Another way to phrase this question is: how do you reshape the data from “long to wide?”
Resources to Reshape Data
The following SAS Note illustrates how to use two TRANSPOSE procedure steps to collapse the observations for multiple BY variables:
If you prefer to use the DATA step rather than PROC TRANSPOSE, the following SAS Note provides a code sample that accomplishes the same purpose:
If the data set is “wide” and you’d like to expand each observation into multiple observations to make it a “long” data set, the following SAS Note is helpful:
Brief Overview of Some Support.sas.com Resources
Since we’ve been discussing SAS Notes from the support.sas.com site, here is a brief overview of how to use the site to find other helpful items for your future coding projects.
- Go to support.sas.com
- On the left side of the page in the Knowledge Base section, select Documentation, Product Index A-Z, and then select your product of interest to access user guides and other documentation resources.
- Go to support.sas.com
- On the left side of the page in the Knowledge Base section, select Samples & SAS Notes.
- From this page, you can type keywords in the Search box to get a list of relevant notes from both categories.
Note: If you’re interested in specific types of notes (for example, samples or problem notes), select the type of note from the choices given on the left side of the page underneath Samples & SAS Notes.
It was a pleasure to work with you in 2015 when you contacted us for assistance. We look forward to another great year in 2016.
If you read my last post, then you know that I’m giving myself the gift of data this holiday season! For me, collecting data on my diet and fitness habits is a gift that just keeps on giving. Although I may not look at all my data sets on a […]
After doing some recent research with IDC®, I got to thinking again about the reasons that organizations of all sizes in all industries are so slow at adopting analytics as part of their ‘business as usual’ operations.
While I have no hard statistics on who is and who isn’t adopting analytics, the research shows that organizations that do leverage analytics are more successful on average than those that don’t. What we need is a new analytics experience, an experience where organizations can:
- Make confident decisions
- Analyze all their data where it exists
- Seize new opportunities with analytics
- Remove restrictions for data scientists
IDC states that “50.6% of Asia Pacific enterprises want to monetize their data in the next 18 months”. Are you one of them or are you going to let your competition get the jump on you?
Big data (or more specifically how to actually gain some sort of competitive advantage from it) is top of mind for forward-looking businesses.
Our research with IDC gives us a few clues on where to head when it comes to the monetization discussion.
In the recent Monetizing Your Data infographic (PDF) created by IDC and SAS, three key approaches to monetizing big data emerged:
- Data decisioning, where insights derived from big data can be used to enhance business processes;
- Data products, where new innovative data products can be created and sold;
- Data partnerships, where organizations sell or share core analytics capabilities with partners.
Organizations that adopt and combine all three key approaches to leverage analytics are twice as likely to outperform their peers1.
If you’re looking to truly create value from the stores of data you have then you need to look at deploying analytics.
1 IDC APEJ Big Data MaturityScape Benchmark Survey 2014 (n=1255) IDC APEJ Big Data Pulse 2014 (n = 854)
Are you missing out when it comes to data monetization? was published on Left of the Date Line.
You've probably heard many times about the fantastic untapped potential of combining online and offline customer data. But relax, I’m going to cut out the fluff and address this matter in a way that makes the idea plausible and its objectives achievable. The reality is that while much has been written about the benefits of online customer intelligence, it far outweighs what’s happening in most organisations today. In fact, considering how beneficial tapping the data can be, I don’t think enough has been written about what types of online customer behaviours should be tracked and how they could be used to create a better customer experience across all touch points.
So where do you begin?
It all starts with what you have decided are the objectives for your digital presence – are they to register, to make a transaction, sign up for a newsletter, interact with a certain content object such as internal or third party? Those are generally the key objectives I see organisations having in order to understand the customer journey leading up to these events, as well as tracking and ‘remembering’ when the customer interacts with all the organisation’s available channels to the market. A key aspect is to monitor and understand how external campaigns, in-site promotions and search contribute towards those goals and how this breaks down into behavioural segments/profiles.
Recognising a customer
The next important consideration is – how do we recognise visitors/customers we should know from previous interactions even if they haven’t identified themselves on this occasion? Identification doesn’t have to be dependent on a log-in. It could be through an email address we can match with a satisfactory level of confidence, or it could be a tracking code coming from another digital channel where customers had earlier identified themselves. It’s of much greater value if we can match their behaviour as unknown visitors when the identify themselves and not have to start building our knowledge from scratch at the time of identification.
This leads to the point where we need to explore our options for weaving a visitors’ online behaviours into our offline knowledge about them and how – at the enterprise level – we can best exploit the capabilities of our broader data-driven marketing eco-system. We should ask ourselves, is it valuable to us to be able to send a follow up email to the ones that abandoned a specific form? Can our call centre colleagues enrich their conversations by knowing which customers downloaded particular content? How important is it to us as an organisation to be able to analyse text from in-site searches and combine it with insights driven of complaint data from our CRM system? What are the attributes of the various parts of the journey leading up to completing an objective?
Perhaps you wonder what I mean by the capabilities of the ‘broader data-driven marketing eco-system’. Well, my point is that it that it puzzles me that most organisations today can’t integrate/report/visualise online customer intelligence in the systems that already comprise the backbone of their information infrastructure. They don’t utilise their existing campaign management systems to make decisions on what’s relevant for the individual and drive online personalisation which increase the online conversion rates, but at the same time can be used across channels. Organisations rarely take ownership of online customer data or use their advanced analytical engines and existing analytical skills to drive next level insights.
Not taking full advantage of campaign management systems already in place is opportunity missed because the deliverables of integrated online and offline customer intelligence are very real. We should be looking for them every day.
This post first appeared on marketingmag.com.au.
It seems like everyone is searching for ‘best practice’ these days. We are constantly looking to learn from what is being held up as good, leading and perhaps even the best itself. While this is a valid exercise, I believe we are missing an opportunity to take a closer look at ‘bad practice’. That’s when either people, processes or technologies create a scenario in which the business case doesn’t hold true and where the project – or parts of the project – eventually fails.
However, we really should take that opportunity because there are important, valuable and very tangible lessons to take away from most such cases.
1. Avoid approaching a one-to-one customer engagement project as a data integration exercise
Building a data-structure that supports the organisation learning about customers’ interactions, responses and preferences over time is a data integration discipline, but the initial approach needs to avoid looking at it that way. It might be that the organisation has 50 different systems with customer and market data in them, but the incremental value of integrating the last 45 of them might not be very significant and only hold very little competitive advantage. I’ve seen organisations spend 12 to 18 months on data preparation and data design but by the time their communication actually hits the channels, things have almost certainly changed. The customer might have new priorities and competitive forces could have shifted.
So my advice is – spend time thoroughly understanding where the valuable use case and competitive differentiation is and build the data processes, the analytics and the automation to address your highest priority use case. Doing so will get to a business outcome much faster. Moreover, it makes it much easier to ask for additional funding to add new data sources, new channels and grow your model’s maturity.
2. Don’t overlook or underestimate how data-driven customer engagement impacts your current way of working
Tailoring emphasis and investment to an analytical way of going to market is easy in theory but hard in practice. Intelligent, real-time recommendations to point-of-sale systems or call centres are only smart if they are being actioned. Call centre workers are not marketers and the churn rate in such teams is often high. So work with them and ask for their input as to how offers and service messages should be served in order to make their everyday life easier. Ask what they think could create a better customer experience in their customer touchpoints. This will not only refine your requirements, but will also start the much required change-management process at an earlier stage.
Take the same approach when aspiring to analytically optimise customer contacts – right message, at the right time – you know the mantra…Recognise that optimising won’t work at all if the business process is designed in a way that has brand managers or branch executives assigned to groups of leads/customers to market themselves to by the beginning of the month or quarter.
3. Don’t focus on functions and features before balancing them against a solid understanding of the implementation team’s skills and experience
Having been through buying cycles, implementation projects and even business-as-usual states a few times now, it always strikes me how much time and effort an organisation will invest in a near-FBI-style interrogation of functions and features when they are choosing systems to drive their multi-channel customer engagements. I’m not saying functions and features aren’t important, but the weight buyers attach to them needs to be balanced against a thorough understanding of the people and skills their vendors can provide in order to deploy and support the software in a timely and high-quality fashion.
As with my ‘overlooking’ and ‘underestimating’ point above – the days are long gone when marketing was just nice to have and ‘so be it’ if campaigns got delayed a little. If marketing is critical in driving tangible sales and customer experience outcomes then systems selection and implementation require a close relationship with the software vendor/system integration partner. This will ensure the business implements the right functions and features it needs within the right time and of the right quality.
The post Avoiding the pitfalls of multi-channel customer engagement appeared first on Left of the Date Line.