Aside

SOCIAL MEDIA AND ADVERTISING

SOCIAL MEDIA AND ADVERTISING

SOCIAL MEDIA AND ADVERTISING

Social media sites can provide valuable information about customers’ views and needs more quickly than traditional forms of feedback. Even the analytics leaders need to recognize that malicious crowdsourcing can distort the results of social media analytics (Guo, 2013).

The three main ways in which unscrupulous companies use crowdsourcing to distort reality are by:

  1. Faking positive or negative reviews
  2. Triggering irrational “herd behavior”
  3. Distorting opinion

 

The competitors are not the only ones who have a motive for posting misleading feedback. Sometimes an organization’s employees may have a reason for doing so too. For example, vehicle manufacturers examine social media feedback to determine the success of their dealerships. As reviews affect the ratings of different dealerships and sales and service teams, employees could post negative feedback about other dealerships and positive feedback about their own (Hamilton, 2014).

(Mattos, 2013).

Procuring or posting fake reviews contravenes corporate policies in many companies. Amazon attempted to prevent it by putting “Amazon Verified Purchase” endorsements next to reviews. However, analytics leaders still draw aggregate reviews — one of the most frequently used filter criteria — from non-verified purchases.

 

Companies can use pre-existing public conceptions or misconceptions to perpetuate a message for their own gain or to distort a competitor’s view of reality. They can do this by (indirectly) taking micro workers to rephrase what real users have been saying about products and companies. It would be very hard for any SMA approach to distinguish between the opinions posted by real contributors and those posted by micro workers. An unscrupulous company should use crowdsourcing to amplify certain problems identified by a competitor’s customers is a great example. This will severely distort the results that the competitor derives from SMA. For instance, the problem reported by real customers could appear in SMA results to be the third most frequently mentioned issue. This could lead to poor decision making by the duped business. Such activity is extremely difficult to detect.

Detractors can use crowdsourcing to post malicious messages, which makes analytics leaders question the value of SMA

Those wishing to post malicious messages can do so quite easily by using a form of crowdsourcing— also referred to as micro work or piecework 8 — via online marketplaces. Hundreds of marketplaces exist, such as Amazon’s Mechanical Turk, Fiverr.com and TaskRabbit.com. The anonymity afforded by the Internet allows users to post malicious messages with little difficulty.

Detractors are more likely to use some channels than others. For instance more use of micro blogging site that is twitter than real-identity-driven networking sites such as LinkedIn. Those wishing to post malicious messages find it faster and easier to set up rogue accounts on micro blogging sites such as Twitter and Sina Weibo. Posts on micro blogging sites are short and quick — readers often take them at face value. Readers also view them in the context of a conversation stream, which allows malicious detractors to make an impact quickly (Guo, 2013).

Recommendations:

Recognize that the results of SMA may not always be reliable.
Ensure that your solution provider understands these issues well and can address them via contributor analysis or can filter analysis by channel source (Hamilton, 2014).

Intentional detractions are hard for analytics leaders to identify, and poor decision-making can result

Many companies present SMA as a way to save time and money on traditional market research. They believe social media sites give customers the opportunity to talk about enterprises and their products in a relaxed way using a medium with which they feel comfortable. Traditional market research feels staged and artificial in comparison.

Where as an unscrupulous business could harm a competitor by intentionally creating one misleading post after another on social media sites. Many false alarms could distract the competitor from dealing with real problems. This has led to the understanding that SMA is not reliable.

Many companies deal with this by conducting traditional market research alongside SMA. They recognize that the results can be misleading if they analyze feedback from either channel in isolation. Combining the results gives them a more reliable picture.

If an individual work in a highly regulated industry, using SMA may not be worth the risk — many companies in such industries have opted to use only traditional feedback channels, such as surveys. Regulations may dictate that the person must take action as a result of negative feedback. Pharmaceutical companies, for example, require to report any adverse effects of drugs. The success of a drug could be severely hindered by acting on the study of intentionally misleading social media posts. R&D cycles could also be affected.

Recommendations:

Employ multichannel analysis and look for consistency in feedback across channels.
Keep records of malicious detraction, whether identified by communities or automated systems.
Share your records of malicious detraction with industry consortia to help alert your peers.
Work with regulators if a person operate in a highly regulated industry to determine what types of social media content are subject to compliance processes.
Work with regulators to develop a way of vetting social media content, which may make the introduction of a reporting process unnecessary.

Benefits and Risks of social media in advertising

Companies benefit from social analytics when they have a clear purpose and use for the analysis in mind. For example, an American news media outlet leveraged social analytics to identify the most shareable format for its content and was able to restructure its advertising formats to attract additional advertising revenue. Importantly, although siloes instances of analyzing social media, such as comparing business sentiment to competitors, can deliver business benefits, maximum impact can be achieved when social analytics integrate with other data sources and analytics processes. For example, when sentiment as added as a dimension to calculate a churn metric, to a customer profile, to a risk profile of a banking loan customer. As an input to fraud detection or homeland security detection or alert detection (Hamilton, 2014).

However, there are risks associated with social analytics. First, the analytics on social data can lead to inaccurate conclusions. One reason is the techniques for accurately measuring sentiment in particular which are still evolving and often involve making a trade-off between time and effort versus accuracy. For example, the larger the sample data size on which algorithms run and crowdsourcing used to assess sentiment the higher the accuracy. However, this is more time-consuming and costly. Making sentiment more accurate with fewer resources is an area of active research, where vendors are investing and on which they are trying to identify themselves (Hamilton, 2014).. Second, governance of data the company does not control affects the accuracy of social analytics. Companies must have ways to identify the quality and integrity of social media — e.g., a competitor or the business itself manipulating social media to influence the customer understanding of analytics and conclusions from social data analysis. Third, real-time requirements would depend on the application/use model. Companies should match their social analytics solutions to their use cases with respect to the type of social media they use and the real-time requirement for analysis to achieve the most accurate results for a particular use case. Twitter analysis for crisis management would require more real- or near-time capabilities than would sentiment analysis of blogs to adjust marketing messaging or campaigns or to conduct time series analysis.

Price Performance

The cost of ownership of a social analytics tool varies greatly from about $18,000 per year to high six-figure deals. This all depends on the scope of the work to be done and the pricing model. Social analytics application providers price in a few different ways. Pricing models with examples are as follows:

Data Size — Prices based on the amount of data that is analysed and comparatively, the number of brand mentions and keyword mentions. This is favorable for companies with a tight scope of work, clear and defined terms and objectives. This is unfavorable for companies looking to scale their social analytics program based on newly discovered insights.
Number of Social Channels — In this model, price is on the number of social channels analysed. This is a favorable approach for a small or midsize business (SMB) or department with a centralized customer team that knows their inventory of social content. However, this is very unfavorable for large enterprises with multiple brands or multiple marketing or sales branches.
Number of Simultaneous Campaigns/Topics Being Analyzed — This is usually a pricing model sitting at a higher price point as it includes the realistic need to bring in new data sources and fluctuation of brand or keyword mentions, so this model tends to be pursued by enterprises and offered by top-level providers. This pricing model is out of reach for SMBs and most midsize enterprises.
Named User — This is a typical way to price plan software, and like pricing by simultaneous campaigns or topics, it is often preferred by large enterprises due to the fluctuating data needs. This is a favorable model for an organization that has a set team of market researchers or business analysts that consume the information and disperse reports throughout the plan. This is unfavorable for companies looking for a social analytics tool to support the business user.

Pursue Metrics That Are Most Appropriate for Your Organization

Just because a metric reflects high business impact for one company, does not mean it will reflect the high business impact for another. This is mainly because social initiatives for CRM are not always focused cross-CRM, but rather on just customer service, just marketing, or just sales. One large media client undertakes a social market research project, where it was saving tens of thousands of dollars per business insight by doing the research through social media versus traditional surveys and focus groups. Aside from the cost savings, its time to business insight was way down, but that was not what business management cared about. Instead, it cared about the ability to sell more advertisements — revenue-generating opportunities, not cost savings. Depending on the business objectives, an individual will have to adjust which metrics an individual seek out and the tactical approach a person take to meet those metrics.

Customer service organizations primarily concerned with cost reduction might look at average handling time for a social media inquiry and at ways to reduce time and cost. However, their marketing and sales counterparts might look at this differently. While both groups are mainly concern with efficiency, marketing may want to spend more time with clients to boost engagement and identify revenue-generating opportunities. Instead of trying to drive down average handling time, marketing might measure the engagement rate (that is, the rate at which the company interacts with clients versus a one-sided conversation). Sales might approach the entire situation differently. Instead of trying to handle every inquiry, it would focus on the ones that present themselves as revenue-generating opportunities and on the number of leads generated through social media.

Hertz’s social customer service team focuses on improving customer sentiment and customer satisfaction in order to create future revenue-generating opportunities. The company has found that by keeping its time to first contact with the customer to less than 60 minutes, it is able to improve customer satisfaction scores and, subsequently, create future opportunities with those customers. As such, agents at Hertz focus on the tangible, controllable metric of time to first contact in order to impact customer satisfaction, with an objective of tracking revenue-generating opportunities and reducing the risk of churn.

The most important metrics for this goal might be the number of questions asked, the number of questions answered, the average handling time, or the number of social media inquiry as a percentage of total inquiry volume.

Customers need to find something different and more favorable about the experience offered by social media support in order to choose that channel for their needs. If customers can wait on hold for 15 minutes, but then have a 90% chance of their problems solved most are willing to do so rather than wait for a response on Twitter that only has a 15% chance of solving their immediate issues. So, customer support on social media needs to be than customer support, it needs to be engaging and differentiated experiences that feels personalized to your customer and develops a rapport that encourages repeat discussion. In that sense, businesses do not want interactions that qualify as “one and done.” Instead, companies want to encourage a dialogue because a dialogue builds a relationship, and relationship and experience are what has customers returning to a person on social media rather than via a phone call. In this scenario, engagement rate, customer satisfaction and the number of social media inquiry as a percentage of total inquiry volume might be your most critical metrics.

Traditional metrics are not always the best metrics to be imposed on social for CRM. Instead, businesses need to consider metrics that contribute to shaping employee and customer behavior, which will then help the company meet its end goals of, for example, cost reduction and risk reduction, and creating revenue-generating opportunities. Consider how one set of metrics, like risk reduction, can help to drive the other set of metrics, like revenue-generating opportunities, and drive the performance we need to meet the objectives.

Measure How Much Top-Down Support There Is for Enterprise Social Networking and Employ Tactics Aligned With Leadership Attitudes

The critical action required by an organization’s senior executives is to mirror the desired new behaviors they want the enterprise social networking initiative to support. If they expect individual contributors to interact with colleagues, share knowledge and communicate openly using social networking products, then they must be the exemplars of this collaborative style of working.

This means that the social networking champion needs to know the leaders’ attitudes about enterprise social networking as an enterprise practice. Readers may believe it is critical for enterprise performance, or that it is a waste of time and money — or may believe somewhere in the middle. Surveys, interviews and research will reveal what these senior executives really think of social networking and whether it will help the enterprise. Surveys, interviews and research can also uncover how much of a change leaders are willing to make in their own behavior. Social networking champions should use these techniques to collect value information they can use in planning the initiative.

Survey. Create a short survey to see how supportive leaders are of enterprise social networking. It may be helpful to conduct an anonymous survey in environments that are politically charged. Collect the results and present the findings to the leadership team. Several potential areas of exploration include senior leadership views on whether:

The corporate culture recognizes and rewards people for contributing to the success of others.
IT investments in traditional and transactional applications can deliver more business value than collaborative applications.
IT investments for corporate growth are more important than those that help the enterprise control its costs.
The IT organization should take a leadership role in helping other parts of the organization change how business gets done.

Interviews- Leaders should be interviewed individually. Collect the results and present the findings to the leadership team. Alternatively, use a focus group to interview many leaders at once and gather collective insight.
Research. It is also useful to examine how leaders are currently using internal or external social tools. Do they participate in communities of practice? Do they have a LinkedIn profile? Do they tweet or blog? Collating this information is an additional source of insight.

Armed with information on leadership attitudes about social networking, the champion can use tactics that are in line with what the executives think. Potential attitudes range from very resistant to enterprise social networking to very supportive.

References:

Afuah, A., & Tucci, C. L. (2013). Value Capture and Crowdsourcing. Academy of Management Review38(3), 457-460.

Chon, Y., Lane, N. D., Kim, Y., Zhao, F., & Cha, H. (2013). A Large-scale Study of Mobile Crowdsourcing with Smartphones for Urban Sensing Applications.

Clough, P., Sanderson, M., Tang, J., Gollins, T., & Warner, A. (2013). Examining the limits of crowdsourcing for relevance assessment. Internet Computing, IEEE17(4), 32-38.

CROWDSOURCING, C. O. (2013). CONCEPTS, THEORIES, AND CASES OF CROWDSOURCING. Crowdsourcing, 1.

Guo, W., Straub, D., & Zhang, P. (2013). The Impact of Formal Controls and Relational Governance on Trust in Crowdsourcing Marketplace: An Empirical Study.

Hamilton, A. (2014). Mass Potential: Exploring crowdsourcing as a tool for public participation in urban planning.

Kontokostas, D., Zaveri, A., Auer, S., & Lehmann, J. (2013). TripleCheckMate: A Tool for Crowdsourcing the Quality Assessment of Linked Data. In Knowledge Engineering and the Semantic Web (pp. 265-272). Springer Berlin Heidelberg.

Mattos, M. (Producer) (2013). why social media advertising[VHS]. Available from http://www.youtube.com/watch?v=rT5MndddX4w

[Print Photo]. Retrieved from http://www.looksmart.com/3-tips-for-increasing-your-roi-in-paid-social-media-advertising

Sui, D. Z., Elwood, S., & Goodchild, M. (2013). Crowdsourcing geographic knowledge. Springer.

Yu, L. (2014). Daren C. Brabham: Crowdsourcing. Genetic Programming and Evolvable Machines, 1-2.