Introduction
Disruptor brands are credited with shaking up previously stagnant industries, but how far does this disruption go? As their large company valuations and customer bases attest, brands like Amazon, Airbnb, Uber, have had major impact on their industries. As with any successful new entry, these new brands get added to consumers’ repertoires, and this affects the marketing metrics, including loyalty metrics, of existing brands/options. While these brands change which brands people buy, do they also affect how people buy brands? Do the new business models disrupt the underlying structure of a category so much that new or adapted laws of growth are necessary? Do disruptor brands, themselves, follow different laws of growth compared to other options in the category?
Science advances through replication and extension, the testing of existing laws in new conditions to identify limits or boundary conditions, outside of which an existing law no longer holds. It is precisely for this purpose that this report tests established laws of brand growth in categories that have been shaken up by disruptor brands. The aim is to see if the act of disruption of the category bends or breaks the marketing laws that have previously held for decades in many other situations.
What might we expect to see? Disruptor brands, through the nature of their disruption, might attract disproportionate loyalty from the customers they attract. These brands also might appeal to a specific type of category buyer, and/or compete differently with other brands in the category.
There are many possible hypotheses, however in the spirit of respecting your time we won’t engage in further speculation at this point, but move to the description of the data and the empirical results. To keep this report succinct, as we only briefly detail each law but refer you to prior reports for detailed background.
Data
These data were collected from 2,004 respondents via an online survey, as no recorded data source collects information across the full range of options needed. Respondents were from an online panel, which means that they are comfortable with the internet, and so have no barriers to using the disrupter options, which typically require some digital competence. The sample was 52% Female, 35% were aged 18-30 years, while 30% were aged 51-70 years, 59% work full time, while income range from 24% with under $35K to 5% over $200K. Respondents were recruited from five cities in the USA: New York City (n=402); Los Angeles (n=425); Chicago (n=408); Houston (n=367) and Philadelphia (n=402). With each city, the transport options were varied to match those available to the respondent.
We collected details on relevant category usage: 90% owned or had easy access to a car, 93% owned a smartphone, and 86% reported listening to music for at least one hour on an average day.
The options within each category
Each category contains a wide range of functionally different options. The results below will display the full range, but each category did include:
- Free and paid options – such as for transport, a bicycle and a taxi
- New technology and legacy options – such as for music, Apple music and a CD
- Expensive and cheap options – such as for accommodation, backpacker hostels and hotels
This range of functionally different options does make this a strong test of the laws of growth. It would take a robust law to hold under these conditions.
Laws of brand loyalty
We first explore two aspects of brand loyalty: Double Jeopardy and sole brand loyalty.
Double Jeopardy
The law of Double Jeopardy states that small brands are small for two reasons, they have far fewer users who are slightly less loyal (see Ehrenberg-Bass Sponsor Report #26). This law is important as it prescribes a predictable level of loyalty and path for brand growth. Typically brands vary much more in penetration than they do in loyalty, and so to grow, brands need to focus on the acquisition of many more customers (see also Ehrenberg-Bass Sponsor Report #31). Strategies that prioritise building loyalty over expanding the customer base are unlikely to be successful (Sharp, 2010).
The results show that the law of Double Jeopardy holds in these categories. For example, in the music category, small brands, such as Tidal and Jango, have many fewer listeners, who listen slightly less frequently, compared to the audience for more popular options such as Youtube or the Radio. Brand penetration varies over 25 fold (57% to 2%) while loyalty varies much less (below 2 fold). Radio is the slight exception, with 46% share of listening, which is higher than it should be given its penetration, but still lower than 50%.
For transport we have separate data for each city, to capture the variation across the five cities (see Table 2 which has results for two of the five cites, New York and Chicago). In all cities, ‘own car’ was the most common option, with the highest penetration and frequency of use, although in New York, the subway is very close in penetration. In line with Double Jeopardy, smaller transport options suffer from many fewer customers, and less frequency of use, though only slightly. Motorbikes, Bicycles and Moped/Scooters do appear to be slightly niched options, with higher frequency than we would expect for their market share – their lower than expected penetration (for their popularity level) being most probably due to lack of ownership and the requirement of a specialist licence locks many people out of these transport options, particularly potential light users.
For accommodation we examined the penetration and the number of times stayed at that form of accommodation over a 4 month period (this included the holiday season with Thanksgiving and Christmas). There is a much larger difference in penetration than loyalty, with an eight-fold difference in penetration with less than twice the difference in loyalty. However some of the paid, lower penetration options did have higher frequency than the free, more popular, options. Perhaps this reflects the adage that guests and fish smell after three days, and people on longer trips are more likely to look for paid accomodation. Share of nights does, however, follow a typical Double Jeopardy pattern.
Therefore in general the law of Double Jeopardy holds, despite the substantial functional differences across options within these categories. The exceptions we do observe reflect some of these functional differences, rather than being prevalent in the disruptor brands.
Sole brand loyals are few, and tend to be lighter category buyers
The next law we examined relates to sole loyals, who are customers who devote their purchases in a category to only one brand/option. Past reports show that 100% loyals are typically a small proportion of any brand’s customer base, and that those who are 100% loyal tend to be solely loyal because they are less frequent buyers (see Ehrenberg-Bass Sponsor Report #35). We find similar results for these categories (Table 4). For the vast majority of options in all three categories, Solely loyals make up less than 10% of the customer base.
It is in transport where sole loyalty is higher. Of own car users, 22% are solely loyal in New York and 31% in Houston. Similar levels were evident for other cities with Los Angeles at 26%; Chicago 25%; and Philadelphia at 35%. However after this option, Solely Loyals drop to below 10% of each option’s customer base.
To assess the potential value of these solely loyal customers, we compared the category usage rate of those who are Solely Loyal for an option, with other customers. Table 5 shows a range of examples for options with higher Sole Loyalty. The results confirm that Solely Loyals tend to be lighter category buyers, using the category 2 to 3 times less often than do non-solely loyals.
Therefore we conclude that just as in other categories, sole loyalty is the rare exception, and tends to coincide with lighter category usage. People tend to be repertoire buyers/users, particularly even more so as they interact more with a category.
Heavy category buyers/users tend to have wider repertoires
Following on from examining the incidence of sole loyals, the next test is whether heavy category buyers tend to have larger repertoires (see Ehrenberg-Bass Sponsor Report #71). A common strategic trap for marketers is to decide to target heavy category buyers and try to make these buyers highly loyal to your brand (see Ehrenberg-Bass Sponsor Report #65). A normal pattern is for a correlation between category usage frequency and repertoire size: People who buy the category more, buy from a larger number of brands. We also see this in these three categories (see Table 6). Heavy category buyers/users tend to use a wider range of music, transport or accommodation options.
Laws of brand competition
In this section we test two laws of brand competition: the Duplication of Purchase law and law of the similarity of Brand User Profiles.
The Duplication of Purchase Law
The Duplication of Purchase Law says that brands share customers with other brands in line with the share of each competitor brand (see Ehrenberg-Bass Sponsor Reports #51 and #53). This law helps us understand market structure, and provides benchmarks for sharing customers with competitor brands. These benchmarks allow us to easily identify partitions, which are where a brand shares more or fewer customers with another brand than it should.
Table 7 shows the duplication results for music, in which we illustrate the Duplication of Purchase Law in a cut-down table of ten options, to make it easier to see the patterns. It shows that of the listeners of music on a radio, 72% also listened to music on YouTube. This sharing figure is similar to listeners of Pandora, of whom 76% also listened to music on YouTube. Every music option’s customer base has around 80% who also listen/watch to music on Youtube, 60% who also listen to music on the radio, down to around 25% who listen/watch music on Vevo.
The bottom row, which has the average sharing figures, shows the overall Duplication of Purchase Law pattern, namely that average sharing of customers declines in line with the brand’s popularity.
For accommodation (Table 8), we find that in general, options share customers with other accommodation options in line with competitor share (the bottom row). Every option’s customer base has more people who have also stayed with family member or in a hotel booked direct with the hotel, and fewer people who have stayed at a backpackers.
However all options do share slightly lower than expected with staying at the home of another family member. A possible explanation for this is that there may be a group of people who have family in another location and a tendency to always visit them on vacation. Others either might not have the family or prefer to spend vacations elsewhere. So this option is slightly partitioned off from others.
Backpackers/hostels do tend to share more customers than expected with several options including vacation rentals and Airbnb. This is common amongst small brands, that tend to attract heavy category buyers. People who use backpackers/hostels are likely more frequent travellers who are more likely to stay in a range of accommodation, including B&Bs, Airbnb and Vacation rentals. It’s also probable that some segmentation underpins this result, with those staying in backpackers/hostels tending to be younger and travelling more.
In the transportation category, the Duplication of Purchase Law again holds (Table 9). Users of any transport option were more likely to also use their own car or Uber, than Lyft or a motorbike.
However we can see a slightly more complex market structure, reflecting the mix of privately owned options (cars, motorbikes, bicycles), cheaper public transport of buses or trains, and the more expensive taxi/rideshare options. For example, those who use their own car are less likely to also use any other transport option (Law of Natural Monopoly as described in How Brands Grow Part 2, Chapter 2), which is linked to its higher sole loyalty. At the other end, users of Motorbikes are more likely to use most other transport options, a common characteristic of a small brand that attracts heavy category buyers.
There is a partition between the two public transport options of bus and sub/way metro, with users of one more likely to also be users of the other.
None of these slight disruptions are surprising, indeed perhaps the surprise is that they are so slight. Of key relevance to our investigation though is that these slight disruptions have nothing to do with the disruptor brands.
Therefore we conclude that the Duplication of Purchase Law generally holds in these categories. While there are some deviations (partitions), these are typically not the disruptor brands, but rather (again) concentrated amongst smaller brands and/or with functional similarities.
Competing brands have similar profiles
The final law we test in this report draws on the finding that the customer profiles of competing brands hardly differ (see Ehrenberg-Bass Sponsor Report #7). For this test, we investigated a number of demographic variables such as gender, age, income and occupation as well as some personal attitudes to innovation1. To illustrate the results we first show the detailed profiles for Music options across Gender, Age and Income variables (Table 10). The brands are organised from most to least popular. The bottom row shows the average across all brands.
We can see the major deviations (highlighted) are concentrated in the legacy technology offerings of Radio and CDs, and in the less popular options (Tidal, Jango, Slacker radio). However these differences should not distract from the overall findings that most music listening options appeal to a wide range of music listeners, particularly the more popular ones. There is little evidence of different offerings appealing to distinct segments of users. A total of 72% of the variations were +/-5pp from the average brand.
In Table 11 we illustrate some of the results for the attitudes to Innovation in the transport category. This was measured as a 5 point scale, but due to the distribution of responses, slightly disagree and strongly disagree categories were merged into a single ‘disagree’ category. The table shows the strongly agree and disagree responses2 for three different statements. The results show some variation across transport options, but this tended to be for the users of functionally different Motorbike and Moped/scooter options, rather than the Disruptor brands. We do not know why around half of both motorbike and moped/scooter users feel they are overly dependent on technology (nearly twice as many as within the user bases of other transport options) – speculation is welcome.
Table 12 shows the Mean Absolute Deviations (MADs) for all three categories (with NYC as an example for Transportation) for the demographics and attitudes to technology and innovation. The results show the average MADs for each category are 4 percentage points. Therefore even though these categories contained quite functionally different offerings, we still see that brand/option user profiles hardly differ. And when it does differ, it tends to be with the small, less popular options, which suggests that differences are less about appealing to a segment and more about not having broad appeal to the masses.
Summary
Our results show that the laws of growth are robust enough to extend to categories with very functionally different options and withstand the introduction of disruptor brands into a category. Disruptor brands do not disrupt the laws of growth for their brands, nor for other brands in the category.
These laws also have obvious implications for marketers of these brands, namely that while you may have a disruptive business model, customers have very well developed ways of dealing with multi-choice situations, which they continue to employ in the face of new options. However we also show that these laws apply to traditional options as well, and so if you are in the situation of having the category disrupted (rather than being the disruptor), then you can still draw on these laws to underpin your competition strategy as you take steps to protect the brand’s market share.
The first important finding is that the evidence points to the way to grow any brand is via expansion of the customer base. And to do this the aim must be to gain many buyers in any given time period (Sharp, 2010). Loyalty metrics will follow. This means a focus should be on mental and physical availability — making the brand easy to be thought of and easy to buy. Brands that disrupt often introduce new innovations in either mental or physical availability. But in many instances there are opportunities for the existing, traditional options to pursue too. For example, Uber’s introduction of online payment has forced the taxi industry to re-look at its payment and booking options; while Airbnb’s promise of a more local experience is encouraging hotels to rethink their offerings to guests. A legacy option/brand needs to adopt or better these innovations to remain competitive. It behooves all brands to constantly monitor whether there are innovations or other approaches to providing their offering, extending their brand’s mental and physical availability in order to build penetration—there is no reason that the taxi industry could not have done some of this disruption of their entrenched practices themselves. And there is time to do this, because when new disrupting options become available, consumers tend to add these to their repertoires, rather than switching wholesale from one option to another. These repertoires existed before the disruptor brands entered, and remain after.
Second, it means that the first users of these new disruptor brands will be heavier category buyers. But to be a big option in a category, a brand/option needs to reach out to the wider category buyer base, including light category buyers – disruptor brands that fail to do this will hardly disrupt their category. This makes activities with mass reach of vital importance to Disruptor brands.
Finally, our results challenge the stereotypes that the users of disruptor brands are radically different from typical category buyers. Its common for pundits to talk about the ‘Airbnb user’ but our results show the Airbnb user is also the hotel user and the backpacker user. We have shown previously that new brands have an ever so slight skew to younger consumers (see Ehrenberg-Bass Sponsor Report #45). Perhaps this is evident here also to a slightly higher degree but this skew is still not substantive and in some cases, such as in music, is partially because the availability of free/lower cost options increases attractiveness to teens with lower or no incomes. Therefore efforts should be put to make these brands even easier to use for older demographics (removing barriers to purchase), who did not grow up in the digital era, but who have the funds to provide an ongoing revenue stream while the teens grow up and gain steady employment.
For the traditional players, the entry of disruptor brands means that their market shares will go down, as the disruptor brands draw share for the existing brands in line with the Duplication of Purchase Law. Some of the purchases will now go to these new brands, while exisiting repertoires and sharing of purchase occasions between legacy brands/options will remain largely in tact. While the new brands are growing substantially, loyalty metrics for existing brands will remain rather stable, including customer retention metrics (and we expect quasi-satisfaction measures such as the Net Promoter score), potentially giving a false sense of security. It will, however, be customer acquisition scores that plummet, which will lead to brand share decline (Riebe et al, 2014).
On a positive note the disruptor options can expand the category, which benefits every option particularly when prior category growth was stagnant.
The laws we investigated hold across categories, countries and time, and so it would be a brave (or foolhardy) marketer who does not seek to understand them and make them integral to any brand’s strategy.
Footnotes
1 Taken from Ratchford and Barnhart (2012)
2 The remainder of responses, to total 100%, were slightly agree or neutral
