An ROI for Social Media Engagement

Social Shake-Up Presentation

The following contains the slides from our recent Social Shake Up presentation. I’ve included both the slides and some speaking notes to help clarify specific point.

Why a Customer Based Valuation Doesn’t Work:

We all want to be able to calculate a value for every customer that reaches out to us via social media. Unfortunately, this isn’t something that’s possible. There are multiple reasons why this is the case, and I’ll cover these in the next few slides. However, if a vendor tells you that they’re able to capture a comprehensive value model for each customer you interact with, trust that they’re wrong.

Your Metrics Baseline:

What you’re reporting on today most likely is based on where your social engagement organization was birthed from. If your team originated within your companies Marketing or Public Relations groups, you most likely rely heavily on tradition metrics from those fields. However, if you started out as part of a call center group then you most likely rely on call center metrics like Average Handle Time. It’s a good idea to be introspective to understand where you started, and where you might need to grow.

It Doesn’t All Mesh:

Click to Enlarge

This is where the bad news starts. There are a ton of metrics across social media. However, they don’t all act the same, and they don’t all interact with each other. Before you do anything else, you need to understand the nuances involved in social reporting.

• Metric Types: Social media platforms have to unique types of reporting, and they are mutually exclusive. The major difference between these two metric categories is:
  • Post Based Metrics (Learn More): Reporting for post based metrics is based on when you as a company complete an action, creating open-ended time windows. This means that the reporting will continue change over time. If I run a report for May on June 5^th, I will get different numbers then if I run the report for May on June 10^th. Each time a customer clicks a link, likes the post, or sees it your metrics will continue to increase.
  • Action Based Metrics (Learn More): Report for action based metrics is based on when a customer takes an action, and has a closed time window. A report for May is the same if you run it on June 5^th or December 5^th. Whenever possible, always defer to Action Based Metric systems. The benefit is that the metrics are fixed, and much easier to understand what happened in a given time period because you don’t have to re-run the numbers over and over. Examples include Adobe and Google Analytics.
• Life of a Tweet: Tweets with self-service links have about a 14 month lifespan. Contrary to that, most social platforms stop recording metrics after 45 to 60 days. These are vastly different time periods. Our research has actually shown the largest volume of self-service link clicks happen after 2-3 months, once the tweet and shortend link becomes indexed by Google.
• Social Metrics: Because of the lifespan of a tweet, we see social metrics like Impressions continue to increase over a 3 month period. We actually see greater than 50% of social activity happens in the 2^nd and 3^rd month.
• Metrics Platforms: Overall, we don’t / can’t rely on a single social CRM tool for metrics. This results in a full value package being developed incorporating inputs from multiple channels. This increases the complexity of the ROI program, but ensures that you develop the best possible picture of what is happening.

Your Metrics Baseline:

Click to Enlarge

If you want to calculate a holistic ROI model for your social media engagement you need to look at all aspects of your work. Most companies looks at their social media activity through a single lens. Either it’s a PR type of activity, or it’s a Call Center type of activity. The reality is, it’s both. When looking for how your social media customer engagement is driving value you need to view it as a combination of a PR Campaign, and Call Center, and a Website. Once you start to look at aspects for all three, you truly start to gain a complete picture of the value and ROI that you’re driving.

The value of social media engagement should be viewed as a combination of a Website, a PR Campaign, and a Call center.

Module Example:

Click to Enlarge

The best way we have found to calculate the value of your social media engagement is to break down the work into individual modules that provide a value to the company. The above module is an example of the process that we went through.

Based on research completed by Twitter, not having an active social engagement process leads to a higher propensity to churn. For example: 36% of telecommunications customers stated they would transfer to a competitor if they did not receive a response from a service representative on social media.

Using this starting point, we created a valuation for a retained customer related to the volume of cases handled. If we, as a company, were not active on social media the cost associated with churn can be demonstrated by the model.

Module Roll-Up:

Click to Enlarge

The above slide demonstrates additional modules that can be leveraged to determine the value associated with your customer care. The modules that are applicable for your company will vary based on your industry and corporate goals. The benefit of this model is that you can combine the applicable modules for the best possible value and ROI. 

Click to Enlarge

Thank you for taking the time to review my slides. If you have any questions please let me know in the comments section.

Advertisement

Automating Web Analytics in Sprinklr

One of the best ways to track the impact from your social media support team is to integrate web analytics into your outbound links. This allows you to combine your social media cases to traffic and actions that happen on your website. It will also provide a comprehensive view of the customer journey throughout the care experience.

Sprinklr, as a platform, has the ability to automate the addition of Web Analytics to every link that is shared by your support agents. (If you use a built in shortener like Bit.ly.) This allows you to automatically tag every outbound link shared by your agents, without the agents having to make any choices. This can be done for any web analytics platform including Adobe, Google, Woopra and many more.

The following instructions cover how to integrate Google Analytics as your automated web analytics platform. These same steps can be followed for Adobe, Woopra or any web analytics platform with slight nuanced changes.

If you’re just getting started with Google Analytics UTM Codes then please visit the Google Campaign Builder. If you’re already comfortable with Google Analytics link tracking, go ahead and skip down to the next step.

Creating a Web Analytics Profile in Sprinklr

Step 1: To get started with setting up your automate UTM link tracking, navigate to the Global Menu in Sprinklr Space by clicking on the 9-dot menu button. Once on the main menu select the “Settings” option from the Social Core section.

Step 2: In the settings menu, select the “Web Analytics” option.

Step 3: Click on “Add Analytics Profile” in the top right corner.

Step 4: Once you are in the build module, you’re ready to set up your web analytics. The system will allow you to integrate both flexible fields and fixed fields. Flexible fields pull in metrics from Sprinklr’s publishing module like time stamps, business units, social media accounts and much more. These allow you to create unique link tagging based on who / what / when / and why links are being posted. Static fields will be something like the word “socialcare” that would indicated in Google that all links are related to Social Customer Care.

In the build module Enter:

• Name [Use something that will be easy to remember and understand later]
• Module Type = Publishing
• Analytics Profile Type = Google Analytics
  • This will automatically populate the possible fields available from Google Analytics.
• Parameter: You can now enter in either “flexible” fields for each option, or static options.
  • Use as many or as few options as you want.
    
  • Each UTM category can have one or more options selected.
  • For Campaign Source: I used four variable metric options,
    and used a “-“ to separate out each value from the next. Each time Sprinklr shortens a link, the system will look at the time-stamp, social network type, account name and business unit associated with the outbound message and add these in the code.
  • For Campaign Medium: I typed in “social-media”. This means that every link sent out will contain social-media as the medium code.

Once you are done, click on the save button to save your new profile.

Applying your Web Analytics for your team.

Web Analytics can be applied at multiple levels within Sprinklr. You can apply them for the entire Client globally, for an individual handle, or for a Campaign. My recommendation is to apply the Web Analytics at a Campaign level.

Note: Web Analytics are based on your domain, so if you operate more than one domain, you will need to repeat the following actions for each domain.

Step 1: Navigate back to the Home Menu and select “Campaigns” under the Planning Section.

Step 2: If you already have a Campaign set up for your Care Group click on the three dots next to the Campaign you want.

• If you don’t have a Campaign set up for your support team(s) now is the time to set one up. Using Campaigns allows you to group all social media metrics for your team’s outbound posts including Likes, Shares, Comments… The Campaign can be set as a default value for your agents. This way the won’t have to pick the Campaign each time they publish.

Step 3: Select the Web Analytics option from the drop down menu.

Step 4: Click on “Add another parameter”

Step 5: Add a Web Analytics profile for each domain you have.

Example: You will need to enter a line for socialservicevalue.com and one for socialservicevalue.net if you own both domains. You can either use a unique Campaign tracking for each domain, or use the same one for all. This will be dictated by how you want to slice & dice the metrics.

• Note: This can sometimes be tricky, and will require a little bit of testing. You can use a * at the beginning and end of the domain to get the coder to be more general. Example: *socialservicevalue.com* will pick up reports.socialservicevalue.com as well as the base domain.

Congratulations, you now have automated Web Analytics set up for your outbound links shared with customers.

Best Practices and Warnings:

• Warning: If you change your UTM codes after a few months of reporting it will confuse your data on the back-end. Put some time upfront into determining what you want to know. This will ensure that your historic metrics will align with what you’re doing in the future.
• Best Practice: I always recommend using the time stamp in your coding. This ensures that every link sent is unique, and you get visibility into the life of your posted links. [We currently see a 14 month lifespan on shared support links.]
• Best Practice: Make sure the variable fields you include in your UTM codes don’t have the option to be blank. Sprinklr has a lot of metrics fields that can be blank from time to time which will also confuse your link data.
• Best Practice: Setting up your Web Analytics at a Campaign level will allow you to separate out link tracking by Call Center or Business Unit. This will allow you to further refine your link and website data to determine which teams are driving what types on interactions on your website.
• Note: For this to work, the link must be shortened by Sprinklr. This can be done using any shortening platform you have integrated into Sprinklr. This can also be set as a default for the Social Media Agents so that they don’t have to select the shortener each time they publish.

Why You Can’t Trust Sprinklr Link Clicks

Within reporting for the Sprinklr Publishing and Engagement modules you will find 8 general link tracking “Metrics”. This doesn’t even include the platform-specific link clicks like “Twitter URL Clicks”. These metrics will allow you to track the user clicks on links that your team shares on social media.  All of these metrics vary in slightly different ways, except the one way that matters the most.

Sprinklr calculates clicks based on link hashes, with no regard for the individual post. When a link is posted, it is tagged with a unique identifier, and this is permanent within the platform. So when the same link is used in multiple posts, there is no way for Sprinklr to know which post it is based off. Because of this, clicks are attributed to the first post, and compound over time.

Note: This is true even when you are auto-tagging links with web analytics, which would theoretically create a unique url each time.

While this might lead to some issues for teams publishing marketing and PR content, it is a major issue for Care groups. This is because care teams tend to share the same links repeatedly. If you think about the 80/20 rule, the majority of your incoming support tickets are going to relate to a few recurring items. These will be thinks like: Reset my password, what deals do you currently have, my billing is wrong, and so on.

Let me demonstrate this with one of the most basic link types that our teams share. When working with customers on Account related issues, we move them from the Twitter public space to the Twitter DM space. THis is done via a direct link to the associated handle’s DM. 

In one month we shared our direct link to Twitter DM 4,867 times. When you look at the example report below you see this same link listed over and over again with over 25k clicks. While anyone would all be ecstatic to have this many clicks on a sales link shared organically in Twitter, we know that this is an absurd value for a Care link. This overstates the click values by millions because we share this link hundreds of time a day, and it is constantly adding up the number of clicks that this link receives.

Side Note: For reasons we still don’t understand, our monthly report has spit out 603 unique click counts, which I’ve been pushing back on Sprinklr about. Based on their explanation I should have 4,867 duplications of the same number since the link hash remains constant.

Synopsis:

Problem: If you are sharing links with your customers through Sprinklr, and then using Sprinklr to track the click rates, your numbers are most likely exaggerated. The more often you are sharing a specific url, the more painful the impact becomes.

Partial Resolution: Your best bet is to always apply web analytics tagging to every link that you share. You can then rely on your web analytics for an accurate count of traffic coming in to your channel. To learn more about how to do this view my blog post here. Note, this will only work for links to domains you own. At the end of the day, links like the direct to DM link in the example can’t be tracked by your web analytics.

If you have any questions on this topic feel free to drop them in the comments section!

The Problem with Social Analytics

For those of you that don’t know me, my first career, and my undergraduate degree, were in engineering. One of the first things that I learned as an engineer is that it takes metrics and analytics to run a process. Over the years I’ve transitioned from engineering, to operations and now into social media. And as I’ve designed social media operations teams and structures, I’ve consistently become more annoyed with the fact that there aren’t better systems available for analytics and live tracking of tickets. It’s impossible to cleanly manage an operational team when you can’t clearly see what is going on.

There are dozens of analytics platform providers out there, but every one of them tends to miss something vital when it comes to reporting. For example, our current provider can’t link Message data with Case data. Since all posts and case management is done through the tool, this seems like a huge miss. In my previous manufacturing role for Owens Corning I managed the implementation of ApenTech for our plant. This system allowed me to see every single thermocouple temperature, chemistry mixture, machine speed or quality test for the entire plant. It had a beautiful layout with clickable drill down points that allowed me to either get a high level view of how everything was going, or look closely at an individual section of the plant. My question has been, why can’t we do this for social? There are metrics in over-abundance for social media. Why can’t we get this organized in a way that allows operational teams to quickly get a pulse on what is happening with their support groups? Every time a platform demonstrates their analytics package to me, I end up seeing this incredibly pretty but altogether useless bubble chart. And if it’s not a bubble chart it’s something else equally useless for operational management. How do I justify more headcount, or a change in operations based on the chat below. What does that even tell me?
So here’s my conundrum: If you’re building out an operational platform that has hundreds of analytic points available, why wouldn’t you use them? If someone could create a live process map to display and track social operations like the one below, enterprise level companies would literally throw money at them. Each node would allow you to click in and get more detail as to what is happening. You’d have filters to set your time constraints, or filter by team. And most importantly, you’d be able to see the whole life cycle of social support in a single view. My only real question is, why hasn’t someone done this yet?

The Impact Of Spanish Closed Captions: Driving Organic YouTube Views

When building customer support ecosystems, more and more companies are embracing video as a key driver of self-service. We recommend that our clients and partners not only distribute their support videos on their own website, but also on a corporate YouTube channel. Currently YouTube is the second largest search engine, and it is owned and built into the largest search engine. This provides you with built-in SEO opportunities that might not exist within your own support pages.

Based on the recommendation that we make to our clients, this article will assume that:

A: You are currently publishing your support videos to a YouTube channel.

B: The majority of your support-video views are organic based, and do not have ad-spend behind them.

Even if you’ve gotten your production costs down to a reasonable $/video (~$1,600 / 3.5 min video), your leadership will still be pushing you to drive the highest ROI possible for your video content. To accomplish this, you need to drive your video views as high as possible. At some point, you will maximize the potential organic views that can be driven through optimization of meta-data, playlists and social sharing. The next logical step is to open up your video content to a new user base.

When talking to our clients, we have found that while many companies are already creating English-language support videos, they do not have the budget or ability to recreate them as Spanish-language videos. This leaves a large percentage of your customers without the ability to easily self-support. Not only this, but studies have shown that Spanish language Americans default to YouTube as their self-support search engine over Google.

If you don’t have the budget for Spanish-language video production, how do you support this customer base with your existing content? The answer is the addition of Spanish Close Captions, and the “Translations” function. While not the perfect customer experience, this will make your videos both searchable, and usable by Spanish speaking customers.

“According to a Nielsen report, the average Hispanic spends more than eight hours viewing online videos every month. That is 1.5 hours more than the U.S. average. Despite having a smaller audience, the engagement on Spanish videos is greater than that of its counterparts. Spanish video content simply seems to “stick” better.” https://digital.gov/2015/09/15/are-videos-in-spanish-the-way-to-go/

A: Spanish Close Captions Allows your customer to turn on the text overlay in their native language.

B: Translations Allows you to add a secondary title and description for the video in Spanish. This essentially doubles the SEO meta data associated with your video.

 

The best part of this implementation is the cost. There are providers like Caption Depot that can produce Spanish subtitles for around $20 a video depending on your volume. Note: if you’re having them produce your English and Spanish CC files at the same time, the cost is even cheaper. If you consider the fact that the average customer contact for a company is around $6, then it only takes 10 views to achieve a 300% ROI. This makes for some very easy business-case math.

After running a trial with a client we found that over the 1st three months of life, videos with Spanish Subtitles averaged 52% more views per month then those without. Also, Spanish subtitles were used significantly more per video than English subtitles.

Summary:

• To increase your potential video audience: Include Spanish Subtitles & Translations to your videos.
• Translations essentially doubles your video meta data.
• Spanish Closed Captions allows you to drive significantly higher organic views of your video content over the first three months of life.
• Spanish language users default to YouTube as the primary search engine for Support content.
• It only takes 10 views per video to attain a 300% ROI on this work.