USPS

Author Details:

• Name: Mary Beth Levin
• Title: Manager, Social Media Strategy and Analytics
• Company: USPS

 

Tell us about you, your company, and your team?

I have two teams under my tent: social intelligence and publishing. The publishing team manages content on our 18 accounts on six different platforms. It also manages the ZONE, the first (and so far only) social media ambassador program in the federal government. The social intelligence team takes a 360-degree look at what is being said about us on social and traditional media. Its purpose is brand protection, crisis mitigation, and a better customer experience.

 

How does your role and your team support your company and its goals?

Our goal is to do good and do better. Unlike many other programs which are based in marketing, we are fortunate to be based in communications in a center of excellence model. This means we are a servant with many masters: HQ communications, field communications, operations, HR, government relations, licensing, marketing and even the Postal Inspection Service.

 

Who were the teams and use cases you put this solution in place for? What challenges was this solution intended to solve?

We wanted to map-out feedback coming to us on social media to the 5-digit zip code level, map-out a variety of specific concerns, and provide real-time as well as longitudinal data. We also needed to protect customer privacy.

When looking at concerns coming to us on our corporate social media sites, we noted specific “customer service hotspots”. We also noted that the types of concerns varied by location. At its very heart, USPS is about logistics. We swim in data. At times we drown in data. With social media, this is yet another set of information competing for people’s time and attention. We needed to provide information in a way that was accessible and compelling to people who are already super busy. Our objective was to make things better for everyone: customers and employees. And we needed to recognize that not everyone loves pie charts and bar graphs.

Social listening software has its own limitations: the platform may limit access in its API, biographical information may be incomplete or outdated, customers may not be posting using their mobile phones (and if they are, geolocation may not be turned on). This is why social listening software will map-out to cities and states and can do so for a limited percentage of social media mentions (about one-third for those platforms which allow access). Listening software will pull in product reviews, but not location reviews. And there is the issue of integrating data from various sources.

USPS uses POS surveys, but these have limitations: they only capture the retail experience, tend to focus on extreme experiences, can’t include photos/videos, and results are rarely shared= outside of a select group.

The teams we involved included communications, social intelligence, social customer response, marketing, operations and government relations.

 

What is a recent win, major innovation, or challenge your team has tackled in the past year that this solution helped you address? Describe the challenge, opportunity, and teams involved. 

We think of social media as the world’s largest focus group. In social listening the customer is driving the agenda of the conversation, telling us what is most important.

We developed an interactive heatmap to illustrate concerns coming to us on social media to the 5-digit zip code level, map-out a variety of specific concerns, and provide real-time as well as longitudinal data. We integrated data from 34 online review sites and our social customer response software (customer concerns coming to us on our corporate social media accounts on FB and X). We incorporated our internal tracking software to show not only the destination address, but where it was last scanned (because issues can occur further upstream). In doing so, we provided a one-stop shopping experience for those wanting to know what was going on nationally and in their neighborhood. Users can select one or more postal areas, districts, zip codes, or subjects (tags). The data goes back to January, 2016 and is updated daily.

USPS Heatmap

 

Because what we were doing was so new, we had to go through an 18-month clearance process to ensure data was secure and the information users were viewing had been scrubbed of personally identifying information.

We knew we were competing for time and attention. And we are cognizant of the fact that our audience is not comprised of data scientists who love excel spreadsheets. We intentionally used a combination of qualitative and quantitative data. People can see a heatmap and recognize that they don’t want the area they work in to show as red! They relate to top ten lists and the personal element of customer quotes, images, and videos the dashboard can show. And seeing how they are doing relative to others and shake them out of complacency with a healthy sense of competition.

Working with the unions, we selected local leaders who would give a short presentation featuring the data of that specific location. This was then tied-in to a larger conversation about social media and its role in greater accountability. The feedback received was “I’ve been working here 20 years and no one has ever explained it to me that way. This makes sense.” Employees recognized that unlike other forms of feedback, this information is publicly posted. We promised the unions that the information would never be used punitively.

These first conversations discussed customer concerns coming to us on our corporate social media sites. In response to feedback, we included online review sites. This information was emailed on a weekly basis.

Information about the heatmap was provided in numerous formal presentations to leadership. Anyone with a USPS email address can access the heatmap. For those using the heatmap, there are weekly office hours, an ops manual, and individual tutorials. Online reviews are shared weekly to those on the ground. Area newsletters share the info monthly. District newsletters highlight results in their localities. There have been close to 100 articles for internal audiences.

Southern Area NewsletterSouthern Area NewsletterWestPac Area NewsletterLINK (national, electronic newsletter)District Newsletter

 

What were the business outcomes and value driven by addressing these challenges? Please share specific results, metrics, and/or KPIs that quantify the change.

Results include the following:

• Sharing the info in short meetings lead to a 54% decrease in customer concerns coming to us on social media.
  • And 18% decrease in ECC (calls to the 800-number)
• Specifically:
  • 31% decrease in WIMP (Where Is My Package)
  • 51% decrease in certified exception
  • 85% decrease in WMM (Where Is My Mail)
  • These changes were stat sig.
• Sharing the review info via weekly emails led to a 35% decrease in customer concerns coming to us on social media.
  • And a 17% increase in online review scores.
• Since USPS spends $58 per ECC and $4 per SCR case, the savings can be significant.
• We have applied for a patent. Completed the one-year preliminary application and filed the final application.

The heatmap received external validation:

• Featured case study by the Social Intelligence Lab
• Presented at the SocialMedia.Org Board Meeting
• Presented at the Ragan and PR Daily Social Media Conference 2024
• PR News 2024 Digital Awards winner for two categories: Social Listening and Employee Engagement
• The Postal Regulatory Commission “applauded the Postal Service for using social listening to determine what people are saying about its brand and services on social media. The social media heat map is an effective tool for visualizing customer comments and categorizing them by concern and geographic area.” It is not often that we receive praise from this entity.

 

What were the progress outcomes driven by addressing these challenges? Please share specific results, metrics, and/or KPIs that quantify the change.

We found additional nontraditional use cases:

• The Office of the Inspector General asked: are there passport concerns worth investigating? (Spoiler: no.) Saved a lot of blood, sweat, and tears since such investigations usually take more than a year.
• Government Relations – are concerns coming to congressional offices anecdotal or indicative? The heatmap has been used to investigate the location and show that the customer’s concerns were not the norm. We have now divided the heatmap into congressional districts.
• Proof of concept that employee engagement scores (high and low) correlate with better customer service and savings.
• Test effectiveness of interventions such as a manga promoting customer service approaches.
• A/B testing for new tracking software. No increase in tracking complaints observed, leading to expanded rollout.
• During natural disasters, we use it to identify local concerns (change of address, closure, theft) to inform operations and social media posts.
• Anticipating PR concerns: e.g. the trend line was going up in Puerto Rico because limited storage led to packages being stored outside. And then it rained. We were able to find additional storage and have a ready answer for the press (who did reach out to us).
• Located customer service hotspots for Operation Santa for targeted training. This program provides presents to 25K children every year. We don’t want anything to go wrong here.
• Determined locations for rollout of products and services such as Informed Delivery (now at 64.9M participants).
• Gemba (site) visits: researching before arriving and giving credit when it’s due. “Hey Erica, I saw that you got a shout-out on social media. Keep up the great work!”

Personally, it is extremely gratifying to get an email from someone on the other side of the country saying, “Thank you so much for this email! We only hear from leadership when things go wrong.” It is not exclusively about improving customer service. It is about creating an environment in which good work and good people are honored.

 

What differentiated the Khoros Communities solution in your ability to achieve/deliver on your goals for the year? What in the solution helped enable it? Were other factors involved? 

There were two things:

1. The flexibility and the sophistication of the tagging system which allows one to code using Boolean or Lucene rules. LLM are still in their infancy and many software lack transparency in their accuracy. I know what I want and the terms my customers are using. When I develop the code myself, I know I’m getting exactly what I am looking for; I don’t have to spend time on learning models.
2. Khoros cut its teeth in customer response. They know how to pull data on the backend to integrate with the new interface we created.

 

What parts of your experience working with Khoros enabled you to address the need or challenge? 

The team understood and appreciated our business case and wholeheartedly supported us in developing something completely new.