Word of mouth is one of the oldest and most powerful marketing channels. Even to this day, some businesses live and die by it. According to Nielsen, 92% of people around the world said they trust brand recommendations from friends and family above all other forms of advertising.

Brands that benefit from WoM the most tend to be from high risk & high purchase involvement categories. 

Despite WoM's well-known value, it's still one of the hardest things to create, measure, and manage. For marketers who can solve this problem, it's a major competitive advantage.

Why is measuring WoM so hard?

1. Word of mouth is very contextual and there is no settled definition for what WoM means.

Everyone can agree on the core principle of word of mouth: people talking about you. This is still too vague of a definition for an accurate and actionable WoM measurement model. "People talking about you" can be a random person name-dropping your brand on Twitter, but it can also be another person telling their friend to buy your brand. Both are considered WoM, but the latter will have a much bigger impact on your revenue. 

We have to get specific. When we do get specific, we then must adopt the measurement methodology to fit our context. This high degree of contextuality makes it hard to create or expect a WoM measurement model everyone can use.

2. Lack of visibility into places where WoM happens.

In the digital world, word of mouth traction for brands predominantly happens in what is known as 'the dark funnel': the company's internal Slack channels, 1-on-1 Zoom calls with colleagues, email, closed industry communities, and so forth. These are the channels marketers cannot access, and thus cannot directly connect their work to WoM outcomes. Without direct access to the WoM data in these channels, any potential WoM measurement model loses a big chunk of its accuracy potential.

As a way to offset this, marketers and analysts will typically rely on proxy data and metrics, as well as leading and lagging indicators in their WoM measurement models.

3. The causal dynamics between WoM events and marketing activities are difficult to understand.

Word of mouth can be caused by a deliberate marketing campaign designed to stimulate conversations about the brand. Word of mouth can also be a byproduct of everything else a brand should do well anyway, such as delivering a product that solves an important problem for the customer or delighting consumers by being very receptive to their complaints.

If marketers are to measure WoM and even replicate its success, they need to be able to point a finger at how WoM happened. Did it happen directly as a result of a single input? Did it happen as a result of the chain of inputs that you did? How long did it take for the WoM effect to take shape? Answering this question in a scalable manner will call for technology and expertise outside of our marketing organizations.

4. Word of mouth data is overly fragmented.

If we look past dark social, there is still an abundance of WoM data at your disposal: in public social media mentions, online reviews, podcast mentions, and much more. Because this data is fragmented, it makes it very time-consuming to collect it (manually), and do so continuously - because the data is created in real-time, your data would become obsolete by the time you're finished collecting it. Even in an instance that you use this method, there is no guarantee that you'll collect as much of the WoM data as you could benefit from. This is both inefficient and ineffective. This problem is solved by platforms like BrandOps, which use AI to continuously collect thousands of digital brand signals (including word of mouth signals!) across all digital channels.

Proposed solutions

The following are summaries of various word-of-mouth measurement models.
If you wish to dive deep into them, we will include links to the sources of these methods, with full credit to their creators.

1. Word of mouth coefficient (by Michael Taylor, Yousuf Bhaijee, and Phil Carter).

The word of mouth coefficient is a quantitative model predicated on the assumption that active users are predictive of new organic users (generated by word of mouth traction). Moving from that basic requirement, analyzing this model in its more complex forms allows you to pinpoint the standalone factors which influence WoM such as seasonality, brand messaging changes, and so forth. This allows you to forecast and influence future WoM traction.

What makes this model great is its flexibility and practicality. The depth of this model can vary from simple to complex, based on your unique needs. This means it can be used by SMEs to measure WoM on a single main customer acquisition channel, as well as by big brands with many channels in their arsenal. You don't need any advanced software to apply this model - a basic Excel spreadsheet coupled with Google Analytics can do just fine!

Michael, Yousuf, and Phil wrote an in-depth article about this model, along with a step-by-step guide on how you can create it to measure your brand's word-of-mouth traction. You can find the article here.

Michael also has a video guide on WoM coefficient, which you can find here.

2) The K-Factor

Most commonly applied in SaaS companies, The K-factor is used to describe the growth rate of websites, apps, or customer bases (not exclusively) as a result of referrals. The formula for measuring The K Factor is K=I*C, where I=number of referrals sent by each customer, and C=percent conversion of each invite (e.g. if one in five invitees convert to new users, then c = .2). Aside from measuring WoM referrals, the additional value of this model is helping you understand who your most valuable customers are.

We mentioned that we can't necessarily track word of mouth with full accuracy, because oftentimes WoM happens in private channels marketers cannot access. This is why the K-Factor interprets word-of-mouth as online brand recommendations, such as "invite a friend" forms. Such forms support cookies and data sharing. This allows brands to track both the person referring the brand and a person who converted as a result of that referral using attribution software.

A common argument against The K-Factor is that sometimes people will refer your brand to a friend but won't fill out a form you can track and ultimately consider in your K-Factor measurement. Some other marketers criticize the K-Factor model because it measures all referrals, rather than "real" word-of-mouth virality (the referrals you get by default). This argument comes from the fact that organizations often spend a decent portion of their marketing budget to incentivize referrals (e.g. invite a friend, get paid). There is a basis in this argument, as one of the key reasons to improve K-Factor is to reduce CAC, not increase it.

Honorable Mention: Word of Mouth Equity (by McKinsey)

Being the most conceptual model on this list with the least amount of supporting data, Word of Mouth equity is listed as an honorable mention for that very reason. WoM Equity framework assumes that the power (or equity) of a brand's WoM traction is a combination of the WoM volume (how often our brand messages are shared by other people) and the impact of said word-of-mouth message. While measuring the volume of the WoM spread is straightforward, the "impact" part of the equation is based on pre-determined criteria as shown in the visual below.

In a perfect world, this framework makes sense as it covers the most relevant grounds for measuring WoM. Where it falls short is in its practicality, because applying this model requires a lot of contextual data and manual data interpretation which most marketers won't have on hand. The factors of this framework work on a self-scoring basis: how much any of these factors contribute to Word of Mouth Equity is determined by you or your team. This highly contextual scoring system makes it hard to form a reference point to other category brands to better understand how strong your WoM equity really is. We think this is a very interesting model nonetheless worth experimenting with. The biggest issue of this model, which is the lack of available data (messages, message senders, channels) can be solved using platforms like BrandOps - and the model can then be made practical.

Read all about the nuances of this model from its source.

The conclusion

Despite being one of the most powerful marketing channels, word-of-mouth still faces a measurement problem. Lack of measurement capability creates risk and as a result, organizations tend to underinvest in it. Measuring WoM is a problem with many sides: collecting enough (and accurate data), developing a sound mathematical backbone behind the models, making the models context-independent, and more. It calls for a chain of solutions - rather than a single stroke. Solving this problem is not only a future task for BrandOps, but also for our entire industry.

With the rise of MarTech platforms like BrandOps and the growing trend of approaching marketing through a scientific lens, we feel very optimistic about the future of WoM measurement. If you enjoyed this blog post, we invite you to suggest alternative WoM measurement models that we didn't mention in this blog.

Thank you for reading. We hope you enjoyed this blog post.
If you want to learn about a better way to measure your brand across all digital channels, schedule a BrandOps demo with our team.