What Net Promoter Scores tell us about Twitter and Facebook

In late 2013, I fielded two paid studies using Google Consumer Surveys to gauge Twitter and Facebook usage and Net Promoter Scores (NPS).

The motivation

1. I was already spending tons of energy working on NPS at Fab — it was our #1 metric. I still had some extra cash from my fantasy baseball winnings and wanted to use the proceeds on something to combine work and play (my girlfriend thought otherwise, but was supportive).
2. Twitter was going public and I wanted to contribute to the conversation.
3. Facebook vs. Twitter is an extremely over-saturated conversation, hugely in need of a fresh take.
4. Net Promoter Score has a lot of momentum with brands and startups, yet there are few public benchmarks available, especially for startups.

Google Trend for “Net Promoter Score”

The methodology

→ The survey was limited to adults 18+ in the U.S. and respondents were only exposed to two questions each for either Facebook or Twitter, not both. In total, I collected over 3,600 responses.

→ A respondent must have used Facebook or Twitter at least weekly on average, over the preceding three months. My intention was to survey only active users, and exclude those that have lapsed or become inactive. I screened out those that did not qualify.

Question #1

→ Users that qualified were then asked the Ultimate Question— “How likely are you to recommend Twitter (or Facebook) to a friend?”

Question #2

The results

Active and Daily Users

While it wasn’t the intention of my research, utilizing Google’s screener provided a nice side benefit…I was able to gauge usage for both platforms. Lets look at two key measures — Penetration and Stickiness.

A) Penetration — Needless to say, finding active users of Facebook was much easier than Twitter. Below are the weekly average users (WAU).

→ 2 in 10 respondents used Twitter weekly (2.3); 6 in 10 respondents used Facebook weekly (5.8)

B) Stickiness — To measure this, we can look at the respondents that use each service daily (DAU) as a % of those using it weekly (WAU).

→ Fewer than 1 in 10 used Twitter daily (0.8); just under 3 in 10 used Facebook daily (2.8)

→ The precise DAU/WAU ratios were 33% for Twitter and 48% for Facebook.

Takeaways

1. I feel good about the Google samples. The 2.5x ratio of Facebook WAUs compared to Twitter is reasonable. Twitter has ~55mm MAUs and FB has ~175mm MAU in the U.S., and the survey was limited to adults 18+ (likely excluding a larger proportion of FB users).
2. Depending on which side of the fence you are on, Twitter either still has tremendous upside or, if you’ve been reading much of the recent press, is not yet and may never be mainstream.
3. Facebook continues to have outstanding stickiness despite increasing concerns of its inevitable demise…

Net Promoter Scores

I had two primary goals at this stage of my analysis for both platforms - understand overall NPS and look at the NPS breakouts by age.

Overall NPS

Look, if you’ve spent any time working with NPS, most benchmarks indicate top-down scores between 50 and 70 are pretty good. Anything above that used to be Amazon territory.

So what do the Twitter and Facebook scores look like?

Twitter = -40. Yes. That’s a negative.

Facebook = -47. Ugh. You must be kidding.

I’m not…

If you are keeping score at home, Twitter Promoters = 21%, Detractors = 62%. Facebook Promoters = 19%, Detractors = 67%.

NPS Breakout by Score (Green=Promoter, Red=Detractor)

Remember, these were users that “claim” to be active at least weekly on the regular.

NPS by age

Twitter — Is, far and way, performing best with the 18-24 year old crowd. The picture still isn’t exactly pretty with a NPS of -22, but the NPS for 18-24 yr olds is nearly 40% better than their next closest group. After a large drop, NPS continues to decrease with each subsequent age group. I.e. The older you are, the less likely you are to recommend Twitter. Makes sense.

Facebook — When it comes to outperformance, there are no standouts. Gen X-ers have the best NPS at -38. On the flipside, there is one negative outlier, and by a large margin. The 25-34 yr old crowd reported an abysmal NPS of -65. Too many baby pics? Maybe. Being the first users of Facebook, perhaps they are just tapped out.

Takeaways

1. There are a lot of fluff pieces written about NPS. This isn’t one of them. Realistically, most NPS scores you see are reported after spending a big chunk of change to have an “expert” measure it or after sending a survey to your user base (useful for operators, but not necessarily what should be reported). A double blind survey of you and your competitors is the most objective read you can get.
2. These scores are just ugly. In reality, it means we are talking about two highly-penetrated consumer products. Fundamentally, the question asks whether or not you would recommend the product to a friend. Well, if most or many of your friends are already using or likely to be using a product, you probably do not need to recommend it. While NPS is considered an indicator of satisfaction and loyalty, it may not be the best tool when you are talking about products with 100s of millions of users. Something to consider. At this stage, the question could stand to use more context.
3. I think the point above is particularly true with Facebook, though I’m less certain with Twitter. Relatively speaking, it’s a company with far fewer users and usage, so it may very well be an indicator that users are not going to recommend it. Maybe they don’t know how to explain it. How do you recommend a product you can’t explain? At this stage, it’s not for everybody, that’s a given. What’s not is who or what will carry this brand into the next decade without better word of mouth?
4. There’s been lots of talk about Facebook and usage with teens. I was not able to get a read on that here. What we can see is that Twitter has much stronger momentum with 18-24 yr olds and Facebook may be losing the 25-34 yr old crowd.
5. NPS scores are all relative. You are either comparing yourselves to a set of competitors and/or you are evaluating performance over time.
6. The key element missing from this analysis (due to costs constraints!) is the critical follow up question - “Why did you give that score?” Rather than guessing why these scores are so damn low, we’d have some pretty strong hypotheses. NPS is only truly actionable when you ask “why?”.

If you are a startup not measuring NPS, you should be considering it. Have I mentioned how important it is to ask, “why?”

Last lesson learned. Next time, do this with less mature social networks. Maybe the likes of Snapchat, Vine, WhatsApp, or even Instagram…