May we suggest...

Buffer Reports

How Much Should Marketing be Responsible for the Happiness of Customers? (Our Answer: More Than You’d Think!)

Buffer Marketing Report

August 2016

Key stats:

  • Buffer signups

    1,697 +33%

  • Buffer for Business trials

    266 -2%

  • Social blog sessions

    1,164,934 +14%

  • Email subscribers

    100,819 +1%

  • NPS

    57 -1%

What are some of the metrics that guide the goals for your marketing team?

Leads, signups, conversion rate, and traffic all seem to be popular ones. Some folks might be interested in social media reach or email subscribers.

These all sound great to focus on and would be fun to chase. (They’ve certainly been on our radar in months past.)

We’re in the midst of taking a slightly different route at Buffer, exploring what it might look like to focus our marketing efforts on a product-specific number: Net Promoter Score (NPS).

I’d love to share more about what this looks like and what we’ve learned so far this quarter in exploring NPS as a key metric for the marketing team. In addition, we’ve been up to some neat things on the blog and with our team structure. I’m excited to share it all. Hope you enjoy! Feel free to AMA (ask me anything) in the comments.

What is NPS? “The one number you need to grow”

NPS began with Andy Taylor, the CEO of Enterprise Rent-A-Car, who came up with a simple, two-question customer survey that clearly showed the health of his various Enterprise Rent-A-Car offices. Further research by the team at Harvard Business Review confirmed the accuracy and success behind this metric, and they took things one step further, simplifying the survey to a single question.

How likely are you to recommend [product] to a friend or colleague?

(Answers are given on a scale of 0-10 with zero being “not likely” and 10 being “very likely.”)

Here’s an example from Groove’s NPS survey.

Groove NPS

Responses are divided into three different buckets: Promoters (score 9-10), Passives (7-8), and Detractors (0-6). In the final calculation, Passives are dropped, and you subtract the percentage of Detractors from the percentage of Promoters to arrive at NPS.

Here’s a neat graphic from CustomerGauge:


Here’s a neat graphic from Team Leader:


We show this question within the Buffer dashboard. It might pop up while you’re filling your queue or customizing your schedule. Then after a customer submits a response, they receive a follow-up question: “What is the most important reason for your score?”

The score combined with the feedback provides a neat mix of qualitative and quantitative data for us to sift through and learn from. Here’s a fun one, for instance:

NPS comment

So what does all this have to do with marketing?

Good question!

I’ll admit, we’re far from the first to consider tying NPS so closely with marketing. One team we take a lot of inspiration from is the marketing team at Slack, led by CMO Bill Macaitas. He believes NPS to be the most important metric for everyone he oversees — a group that includes marketing, sales, and customer success. In a feature at Mashable, Bill explained a bit more about his thinking with NPS:

Customer experience and growth should not be viewed as competing forces or a zero sum game. A great customer experience will yield happy customers who recommend you more. Additionally, word of mouth recommendations are usually the best converting lead sources.

Marketing has a huge role in that experience and is often one of the first touch points that a customer has with your brand. It is critical that they have a great first experience.

In another interview, he went into greater detail. The Optimizely blog took some notes.

Bill is not fully satisfied if someone signs up for Slack. He is not fully satisfied is someone becomes a paying customer. Their bar is will someone recommend Slack? “And that’s a much higher bar.”

Holistically focusing on NPS and CSAT (customer satisfaction) are long-term metrics, leading indicators of long-term growth. Leads, opportunities are short-term metrics.

We aspire to be customer-focused to a similar extent: wholly, foundationally, with our hair on fire to see our customers succeed.

And so we arrived at NPS.

From a Buffer perspective, this idea has fit really well with how we’ve historically approached content. For years now, we’ve been interested in delivering a great experience for everyone who interacts with our brand.

We feel responsible for every online interaction someone has with our brand.

We feel responsible for every online interaction someone has with our brand.

And furthermore, we feel that every online interaction feeds into the NPS score.

If a customer interacts with us on Twitter, reads a handful of blog posts, watches a Facebook Live session, then hops into their Buffer dashboard to take the NPS survey, our marketing efforts are going to have an impact.

Even those days, weeks, months before someone joins Buffer as a customer, we feel that the experience someone has with our content will plant a seed in that person’s mind of what kind of brand Buffer is and what level of confidence they can have in our product.

MailChimp has this neat way of explaining why they do so much brand marketing. They want to create a bias for MailChimp so that when someone is ready to choose an email marketing platform, MailChimp will be top of mind.

It’s similar for what we hope to achieve at Buffer. We want to deliver the absolute best experiences with our blog, our social media, our videos, our swag — everything! — so that these experiences create a bias for Buffer and create trust and joy with the Buffer brand.

That’s the high-level, hypothetical stuff at least. (How does it sound to you?)

Now for the specifics.

We’ve not fully turned over all our metrics to be NPS-driven. You can peek at our marketing OKRs here to see how they are still mostly, traditionally focused on traffic and leads.

But what we have kicked off is some deep exploration into what an NPS push might look like in the final months of the year. This exploration involves tracking NPS closely, both the quantitative and qualitative results, and putting together cohort analyses on the effect of marketing on NPS.

Here’s what we found so far:

  • NPS for customers who signed up after viewing any Buffer blog post: 68
  • NPS for customers who signed up direct from the Buffer blog: 67
  • Customers who have opted in to our product newsletters: 60
  • Users who have opted out of our product newsletters: 49
  • Users who subscribe to our MailChimp emails: 64
  • NPS for customers who have subscribed to the Actionable Strategies email course: 69
  • NPS by referral channel (how people came to sign up for Buffer): see image below

NPS by referral source

One of the takeaways we’ve had from this is that email marketing might not be as strong of a channel as we had previously thought (I’ve been on the edge of pushing more resources in this direction, though now I might take a step back). And content like the Actionable Strategies email course seems to really resonate with people.

Beyond tracking and analyzing NPS, we’ve also found NPS to be an inspiring source of products and ideas for us this quarter. Here are just a small handful of things we’ve been able to think about:

  • Offer a referral program to Promoters
  • Mine the qualitative results to find content gaps
  • Starting a customer evangelism program aimed at Promoters

And (maybe as a sign that we have too much NPS on the mind) :) we’ve even found the score seeping into other areas of marketing and Buffer.

  • On the Buffer Social blog (“Would you recommend the Buffer Social blog to a friend?”) — 89 score
  • After our in-person meetups and events (“Would you recommend this event?”)
  • The app we built internally for our Buffer retreat in Hawaii  (“Would you recommend this app?”)
  • Our very team!

It’s been a really fun time being so NPS-focused these past couple months. In terms of next steps, we’re going to revisit all our NPS learnings from this quarter when we set our plan for Q4 — and we might just find ourselves diving even deeper into NPS!

What has your experience been like with NPS? I know I’ve got a lot to learn on the topic. I’d love any of your stories or articles!

(HT to Wootric for the NPS scores. If you’re keen to track NPS for your product, they’re a great solution.)

August’s highlights from Buffer marketing

1. Going from 1.1 million blog visits to 1.3 million

Since the start of the year, we’ve been eager to grow Buffer Social blog sessions from 1.1 million to … well, anything higher. To be honest, things felt a bit stuck.

Until this month.

Huge credit to our social blog editor and lead writer Ash for cracking the code and finding some keys to growth on the blog. This was the number just last weekend. (It’s for the previous 30 days.)

sessions on the Buffer social blog


We did it! And we’re excited to reach for the next milestone.

As is often the case with numbers like these, it wasn’t the impact of any one particular thing that drove us to 1.3 million sessions. We think it was the combination of several:

We’ve been emailing our main Buffer customer list once every three weeks with a new blog post.

We’re lucky to have a list of 1.4 million customers whom we can email on occasion. Typically we send new blog posts to just our blog subscribers. Now, we send the big ones to customers, too, when we feel they’re relevant.

email boost traffic

We’ve been revisiting past blog posts to update them with new info and publish dates.

This historic optimization (something that other sites like HubSpot and Sprout Social have done successfully) has been a great boost to the search traffic on these pages.

Here’s one of those posts:


We’ve been writing more long-form, big huge guides on social media topics.

In the past few weeks, we’ve published a huge Instagram marketing guide and one for Facebook ads. In the coming weeks, we’re hoping to publish about content curation and coding for marketers.

The Complete, Always-Updated Guide to Facebook Advertising

Sidenote: How to measure just some of your blog traffic

At one point in the past few weeks, I was about ready to kick off a significant brainstorm to solve our blog traffic challenge. Then our co-founder Leo reminded me about the type of content we have on our blog.

The majority of our traffic is to our social media content.

A significant percentage is also to our lifehacking content, which we published lots of a couple years ago and have pivoted away from since.

What if … the traffic to our lifehacking content was dipping while the social media content was on the rise? What if I was getting flustered for nothing?

There was only one way to find out. Into Google Analytics I went!

Within Google Analytics, you can set custom segments that filter the data just the way you want. So I set about adding segments just for lifehacking content and just for social media content. Here’s what one looks like:

GA segment

Turns out the lifehacking content represented 10 percent of our traffic and was down about 20,000 sessions year-over-year — a small number in the grand scheme of things but still a number I was grateful to know.

2. Marketing team gets an engineer

Late in August, we were so happy to welcome Steve, a Buffer engineer, to our marketing team as our full-time marketing engineer!

Steve joined Buffer as a designer and has worked on successful projects like the social media calendar and the Buffer style guide. Lately as Steve has embraced a generalist role and developed more of his skills and interests, he’s taken on loads of responsibilities as an engineer, working with the Buffer product and growth teams. Now, his next transition is to join us on the marketing side to lead all of our engineering projects and experiments.

We’ve already got quite the Trello board going for him.

Marketing Engineer Trello board

3. Weekly brainstorm

One of our OKRs this quarter was a simple (and fun) one: meet together once-a-week to brainstorm.

Brainstorming is one of the oft-spoken-about challenges of remote work. When you’re not bumping into each other in the hallways with ad hoc conversations, the opportunities for brainstorming are limited.

So we hoped to create some of our own.

We meet every Friday for 30 minutes to just bounce off ideas and seek feedback on some fun stuff we’re thinking about, working on, or dreaming toward. It’s been great to connect in a more laidback fashion toward the end of the week (our weekly standup is on Mondays).


Looking ahead to September

We have a few big things planned for Buffer marketing in August. Here’s a sneak peek:

  • Launching a social media podcast!
  • Partnering with for a social media webinar
  • Working on a new marketing project — a mini-product that I’m excited to share about next month

What can we share that would be helpful for you?

If you have any questions at all about our marketing plans or projects, I’d be so happy to chat with you here in the comments. Feel free to share any thoughts or questions, and I’ll jump right in!

Check out more reports from August 2016:

  • Lovely article Kevan! So happy to see that Buffer are using NPS. Buffer has always seemed to me to be the quintessential NPS user! I’m a big fan and have implemented it for many clients when I was working for Bain & Company.

    And you are right! The idea of NPS came from Enterprise Rent a Car’s CEO (think he called it ESQI). However, the trademark for NPS belongs to Fred Reichheld, Bain and Satmetrix, who did the research and statistical analysis supporting the Net Promoter Score.

    NPS was also renamed from Net Promoter Score to Net Promoter System. Over the years, after implementing NPS for many companies, they came to realise that the NPS score is only half the story. “The Ultimate Question” requires “The Ultimate Follow Up Question”.
    It is the follow up, open-ended question that makes NPS so powerful because it allows a company to get direct feedback from a customer and act on that feedback immediately. Leading NPS companies all have tight feedback loops between NPS survey responses and the relevant operating divisions. The simplicity of NPS makes it easy for every employee to understand – from back-end to front-line staff i.e. to convert detractors and create more promoters. Many companies also use an eNPS – an employee-NPS. We have research to show that happy, loyal employees = happy, loyal customers. Kinda like Buffer :)

    There are also many detractors of NPS. Many clients find the question indirect and awkward (especially in Asia and non-English speaking countries). Some find the scoring system arbitrary and subject to cultural bias (e.g. the Japanese seldom give 9 or 10 so NPS scores in Japan is always lower). One of the big watch-outs is selection bias – if you are calculating NPS based on voluntary responses, the results are naturally biased as those who are neutral or detractors will not bother to take part in the survey. Low N is another watch out (e.g. N for your email responses looks too low to be statistically significant). Ease of response is another factor; every additional step required will lower your response rates and potentially your scores too (e.g. how many steps does your Mailchimp subscriber have to take vs your blog subscriber vs a Medium reader?)

    Cross company comparisons are not advised unless you are the one administering the survey because the slightest tweak in question or how the survey is administered can make a big difference in results. The other thing to note is the difference between customer satisfaction (which is what ESQi is based on) and customer loyalty. At the heart of NPS is the concept of customer loyalty and loyalty economics (repeat purchases, larger share of wallet and purchase size, lower cost of acquisition). A satisfied customer may not necessarily be a loyal one.

    Digital factors are also affecting NPS, and the idea of loyalty is being redefined. In today’s digital, mobile first, omni-channel environment, it is not uncommon for a person to be loyal to a company, yet purchase from multiple retailers. Social media amplifies the cost of a detractor while influencer marketing can reduce the authenticity and value of promoters. These are all changing the statistical relationships established in the past.

    I could go on and on! The Net Promoter System website and podcast has lots of great resources. I love the Trailblazer Video Series (especially the Zappos one).

    There is a long list of companies who use NPS here:

    Feel free to ping me if you guys have any questions or want more information. Am a big Buffer promoter. :)

    • Wow, thank you so much for this amazing background, Debbie! Your level of detail and history is incredible!

      Really interesting to hear all the pros/cons about NPS, too. This has been some great food for thought for us. Can’t thank you enough for sharing. :)

    • Just curious, Debbie, given some of the challenges with NPS that you’ve identified here, how might you advise us on proceeding? Or what have you done personally?

      • Hey Kevan. There are three types of NPS surveys that we use:

        1. Top Down (or Relative NPS) calculates NPS for the relevant competitors in your industry. It involves creating a sample of customers that are representative of the market, not your company. The most rigorous surveys are double-blind by design to minimize bias where Customers remain anonymous and Researchers don’t reveal who’s sponsoring the survey.

        2. Relationship NPS: Conducted at major touch points in the customer relationship. In complex / long life cycle businesses, we focus on “moment of truth” touch points – inflexion points in the customer life cycle that have a big impact on creating promoters or detractors. This helps to help prioritize which touch points to focus on.

        3.Transactional NPS (e.g surveys conducted immediately after a sale).

        With open-ended feedback, we drill down to analyse the underlying drivers of root cause. Feedback is clustered into categories (marketing, product, customer service) and subcategories (marketing: email course, blog, newsletter; product: ease of use, features etc). The relevant clusters and categories are agreed by the team to see what makes the most sense from a business perspective. We then use text analytics software to do the categorisation to speed up the process. We sum the positives and negatives for each sub-category and hand over the verbatim feedback to the relevant departments (e.g. product development, customers service) who have coached on what to do with the feedback. This creates the high velocity, feedback loop that makes NPS so powerful.

        Coming back to how Buffer uses NPS: in the Referral Channel in your table, are you using a Multi-Channel Attribution Model to attribute Source of Signups? There will be multiple interactions before someone signs up to use Buffer but your table looks like last interaction / last click attribution (which always overstate Direct as the source of conversion). That may be why Email and Social look understated. You’d want to use a Multi-Channel Attribution Model. Bear in mind that the attribution model that you choose (First Click, Last Click, Linear, Time Decay, Position-based) will result in different NPS scores, so pick carefully.

        I’ve never worked on NPS with non paying users before, so I’m not sure what the right answer here is, but instinctively I would ignore all non-paying users of Buffer for NPS purposes.

        For Marketing, I suspect the root cause of someone being a Buffer promoter / detractor is likely go deeper than just being a blog reader or email subscriber. You’d want to dig into the different content type (e.g. culture/values vs product updates) and see if the score changes over time / exposure to the content). You might want to test to see if resonance with Buffer Values or “Transparency” is a driver of high NPS scores.

        For your Email Course, are you looking at those who subscribe or those who actually clicked through and completed the entire course? You want to establish if the email course was effective in converting Neutral customers into Promoters (it might be that it only attracts customers who were already Promoters). Next time you do it, test for NPS scores before and after.

        One challenge I foresee is that a large percent of your open-ended responses will not tell you anything about root causes. (Even answers as fabulous as the one you posted). So you’ll need to find a way to force customers to give more actionable feedback. One way could be to add a second multiple choice question (“Which category is the main reason for giving your score: product, customer service…”) and then have the follow up question (“What is the most important reason”).

        We often add a third question to ask if we can follow up with a phone call to drill down on some answers, and to get testimonials. For your most valuable customers, you could even get Joel or Leo to call personally – definitely a way to make a customer feel valued.

        If you are collecting NPS scores on a regular basis from all your customers, any changes in individual scores could be early warning signals especially for Buffer Business customers, so you might want to push out an online survey on a quarterly / bi-annual basis vs leaving it up to chance – if the person happens to be at the Buffer Dashboard and if the survey happens to pop-up and if they happen to have the time to answer.

        One key difference that I’ve noticed between Bain NPS surveys and others is that we always start by establishing Customer Economics and Customer Lifetime Value (CLV) to understanding how differences in behaviours of Detractors vs Promoters affect CLV. E.g. How often do Detractors churn and how does that impact CLV? Or is the value of Promoters due to their higher Spend or due to Word of Mouth? How much do you save in cost of acquiring a customer that came through word of mouth referral?

        If you don’t know the value of converting 10% of detractors to promoters, you won’t know how much to invest to do so. If you don’t understand what drives CLV in your business, you won’t know where to invest – be it improving retention or upselling or encouraging WOM. Finally a 30% NPS score that comes from 60% promoters, 10% neutral, 30% detractors will have different implications for company that has a 30% NPS score that comes from 35% promoters, 60% neutral, 5% detractors.

        Yikes super long answer but I hope this helps! I’m not the NPS expert and how it is used varies from company to company. But I know for sure that the companies that use it effectively to drive loyalty have outperformed their competitors.

        I’d definitely take advantage of Buffer’s high NPS scores and community of promoters! It would be a waste not to!

  • Our team at CareerFoundry uses NPS amongst all our customer facing teams, but I’m sure we can improve other teams just as well, inspiring! I do have to agree with Debbie, the follow up question gives us huge amounts of data we distill and use as a basis for UX research, it’s extremely valuable! (Although with >1M users might be a bit too much to chew…)

    Talking about inspiration, we’ve adapted a lot of ‘transparancy’ concepts from Buffer as well, and have some exciting findings there too. I’ll be happy to share if interested!

    • Indeed! Great point, Kees. The followup question is such a useful part of it all. The potential impact is worth it to dig through all the qualitative gems. :)

      Would love to hear more about your transparency efforts!

      • I’m sure our content team would be happy to work with you on that – if that’s something Buffer is also striving to (so sharing others stories on that) – can I get you in touch with our blog editors?

        • Thanks, Kees. We’re not quite in a position just yet to do much reporting or partnering on it, though we’re always really grateful to learn from what you and others are doing. Any links or resources would be super neat to see – feel free to send to kevan @

          • Thanks Kevan, if we decide to publish something on that I’ll definitely let you know! keep up the great work on the buffer blog, it has been eye-opening and deeply inspirational!

    • Hi Kees! Yes the volume can be overwhelming! We often use a representative sample. We also focus on high value customers first. In my answer below, I’ve elaborated on how we use clusters and categorisation and text analysis software to help speed up the processing of feedback!

  • Sandra McCann

    Would love to get updates on how the weekly brainstorming is going! That’s a big trouble spot for remote teams (and a frequent excuse for corporate policies that mandate butts in seats in cubes every day).

    • Thanks so much, Sandra! Brainstorming is going great! It’s quite open-ended at the moment: there’s no agenda or plan for the session, it all happens organically. What I’ve been really impressed with is that we’re able to tie some of the ideas into action/execution already, which I know can be a tricky part of it! Happy to share any more specifics here with you :)

80,000+ social media marketers trust Buffer

See all case studies