Big Data: Security, Privacy and Ethics

This weeks reading material explored security, privacy and ethical issues surrounding people’s online data.

Two articles discussed the issue of potential employers requiring candidates login to their Facebook profiles or provide the password. I agree that requesting someone’s Facebook password crosses the lines of privacy violation because with a password, someone can see personal communication that is not available to the public. It’s nice to know that Facebook is taking a stance against employers that do this, as explained in’s article Can Employers Legally Ask You for Your Facebook Password When You Apply for a Job?  It is their responsibility  to protect user information and privacy and they already skate the line on this issue when it comes to using user information for advertisement targeting.

That being said, we’ve been hearing for a long time about the practice of employers making people login to their social profiles and of social networks using user information for marketing. With this knowledge, people should be cautious not to post anything online that they wouldn’t want everyone to see. It’s common sense, really. My company recently required everyone attend a HR education webinar on Social Media etiquette and your privacy rights as they relate to your professional life. The webinar reiterated a lot of the things the articles from the reading said, so it was nice to know that my company opposes asking potential and current employees for access to their social media profiles. I thought it was also a good idea to educate people on what is and isn’t acceptable social-media etiquette so that every employee has a clear understanding of what is expected since this is such new territory.

As far as Instagram repurposing user pictures for advertisements, I’m not surprised. First of all, because Facebook bought Instagram. Secondly, people can’t think that services like this are offered to them for nothing in exchange. That’s not how business works. If the service is free, the company has to be making an income somehow. It just emphasizes the point to only post what you wouldn’t have a problem with the whole world seeing on the internet.

Another interesting perspective on user data is from health-related data. I do think that we should use the technology to help patients stay up-to-date with their health. Empowering patients to be proactive about their health is essential to improve their conditions in many cases. However, as explained in WSJ’s Health Gadgets Test Privacy-Law Limits, the software and tools to interpret the data to patients in a way they can understand is not yet developed. Having worked in the healthcare communication industry, specifically for technologic devices, I agree that patients should not be provided raw data without context or explanation because most of them don’t know how to interpret it. Not to imply they are dumb. The fact is the majority of them do not trained to evaluate the data. It is the responsibility of healthcare providers and organizations to use technology to help educate, care for and make healthcare more accessible to their patients.

The ability to collect real-time data about these conditions is just expediting the process of scientific and medical research that normally involves a lot of funding and official studies and analysis. The automated collection allows us to collect far greater amounts of data, faster, and cuts down the analysis time drastically. I am interested to see how this helps progress medical knowledge and treatment.


  1. Would you have a problem with a potential employer looking at your Facebook?
  2. Have you ever been asked to share your Facebook with an interviewer or employer? (I haven’t… but my company is very against that practice.)
  3. How much access do you think patients should have to their data?

Using Information Aggregators and Data for Online Marketing

This week’s reading From Information to Audiences: The Emerging Marketing Data Use Cases, A Winterberry Group White Paper was particularly interesting to me because it describes the environment surrounding my current job, as an campaign manager at an agency trading desk. I manage real-time-bidded (RTB) campaigns. We also call it programmatic advertising. Programmatic refers to setting different tools to automate the online marketing process.

I use demand-side platforms (DSPs) to create, automate, run and optimize the campaigns I manage. In these DSPs, you create campaigns using different tactics. The main targeting tactics are predictive, behavioral, contextual, and retargeting. Predictive targeting has minimal parameters with a goal of building your audience pool, and you usually bid lowest for this tactic. Behavioral usually involves targeting different data segments, like the white paper explains. Because the campaigns I manage are for very large brands, data providers actually build custom data segments based on what the brand’s planning team decides will be best. Contextual targeting refers to placing ads based on the content of the page. This can be done through keyword lists, site lists. There are also companies that provide content “concepts” to use at an additional cost. Retargeting can be based off of site visits, ad clicks, or other specific actions you program it to target off of. We often run different creatives for the different tactics, and sometimes test multiple creatives per tactic to see what performs better.

You dictate all of this in the DSP, and make sure everything needed outside of the DSP is set up, as well. For example, you must work with agency counterparts to ensure set up remarketing lists, conversion beacons, custom data segments and creative creation. If these things aren’t set up properly, your tactics will not perform as they’re supposed to.

As the campaign runs, you monitor it for performance, budget pacing, and make optimizations to improve overall performance. We do see that programmatic/RTB advertising is more efficient and effective for clients. It allows us to reach audiences more likely to act, on a large scale of inventory, at lower prices than traditional digital ad space buying.

Also, this white paper was written in 2012, and this movement has progressed rapidly since then and continues to do so. 10 Stats That Prove Programmatic Buying Has Become Critical provides a lot of insight into the future of programmatic buying.

As far as the reliability of the data, that’s something that is a constant priority and investment, from all parties concerned. From the results we have seen, it is definitely a significant improvement in effectiveness and efficiency, which I think shows in the directives to move more media spend to programmatic in the future.

Also, I’ve read people’s concerns about the massive data collection, however, all of the data collectors and aggregators are held to very high security and accountability standards, and there are legal consequences for violations. Also, if people are concerned about it, they should keep in mind that traditional data collection was all personally identifiable (name, address, phone number, etc.) and stored in large databases, so one could argue that this is technically more safe.

Message Testing

Both articles from this week’s reading explore the concept of message testing to determine how different strategies and tactics work across different audiences.  Online consumer behavior: Comparing Canadian and Chinese website visitors studied and compared how Canadian Students and Chinese Students (both in Canada) and how different emotional factors influenced how the two demographic groups interacted with and preferred their online information.

Developing Media Interventions to Reduce Household Sugar-Sweetened Beverage Consumption was about a public health campaign to educate the Philadelphia area about the dangers or high sugar consumption in sweetened beverages. This campaign was started with a study of what kinds of messages were most effective for their target audience and then built the campaign based on the study’s findings. This study and campaign were not focused on internet user behavior, where the first one was.

I believe that understanding your target audience in every way possible is the key to persuasive communication efforts, whether it’s advertising, public interest, or otherwise. That way you know how to craft your message to obtain the desired behavior from your audience. This is also true for political campaigns. I think that’s a good example of where a lot of money, talent and effort goes into researching target audiences and addressing what they want to hear. For example, Barack Obama is on social media now…. Why? Because the entire nation is on social media. It’s to reach and engage followers because that appeals to them.

Studying an audience and testing the outcomes of different messages is valuable because you understand how they view your issue, which makes it easier to identify opportunities and challenges and how to approach them. Also, media planners or researches could be incorrect about what messages will be most effective, so it’s important to test them. This can provide insights on how to further improve messaging and audience understanding.  For example, the Sugar-Sweetened Beverage study found that people were less likely to respond to images of obese children than other messaging tactics. It is important to understand and confirm these things to maximize the effectiveness of the campaign.

You can also apply this to advertising campaigns by running tests of different kinds of copy, creatives, media inventory, etc. to see what works best for the current campaign. Clients usually find those insights helpful, too.


1) How much do you research target audiences in your current role?

2) If so, do you think your organization puts too much or too little priority on audience understanding?

3) Do you have any experience with studies or campaigns that use split-testing or message testing as part of their efforts?