Friday, May 03, 2019

Legal Applications of Marketing Theory, part 2

Dominique Hanssens, Natalie Mizik, & Lorenzo Michelozzi, UCLA Anderson, Univ of Washington, Cornerstone Research, Brand Value, Marketing Spending and Royalty Rates

Brands can create value, but require ongoing marketing. When licensed, disputes can arise about royalty rates and marketing support/maintenance effort. Can we measure the financial value of a brand? Quantify relative impact of brand and ongoing marketing spending?

Lots of effort goes into valuing brands. Five major brand value entitities: Interbrand (top 100); Millward Brown, Brand Finance (top 500), Forbes, Eurobrand (top 100 corporate parents). What methods? Only two major methodologies for economic use value. (1) Earnings split: NPV of brand-specific earnings of the business.  (2) Relief from royalty: NPV of notional licensing fee that a company would have to pay for the use of brand.  In principle, both should yield approximately the same estimate. Significant differences between the various list reflect differences in assumptions of respective models, not differences in methodology.

In the top 30, only 5 brands are common in the top 30 to all sources in 4 2018 valuations.  Amazon, Apple, & Google occupy the top spots.  Only 35 brands are common to all 4 lists. Amazon’s valuation varies from 70-150 billion.  Aggregate value of 100 brands on Millward Brown’s list $4.83 trillion is more than 2.2x Interbrand’s $2.02 trillion.  There’s not even consensus on direction of change in brand value for the majority—18 of 35—common brands featured on all 4 lists. 

Why is valuation difficult? Brands have multiple effects on different groups.  Consumers aren’t the only audience.  Employees (some evidence of motivation/effect on firm costs), investors (cost of capital), regulators (better contracts, concessions on infrastructure), competitors, business partners.  Demand-shifting effects and cost-shifting effects.

Pessimistic view of current state of art.  What can be done?

IRS is very interested in brand valuation, especially w/US companies with overseas affiliates that pay royalties into the US.

Brands require constant marketing investment.  They may have different returns to marketing investment—A brand that has to keep marketing to generate sales has a weaker “brand effect” than one that would not lose many sales if it stopped marketing.  Probability of choice w/o marketing stimulus: brand effect.  Helps us understand which sales come from marketing versus from brand. 

Q: do you equate brand and product?  Do they have to be same quality?

A: that’s another variable; have to control for that.  Case-specific.  If it’s product level you have to do it for each product separately. A company may rely more on marketing for one division than for another, or in South America than in Europe.  Product effects tend to be heterogeneous.  Nivea is bottom tier face cream in France, higher in other countries. The rest is all accounting; base royalty rates on that part of the revenue generation that can be attributed to the brand. 

Extensions: brand might not just influence final consumers.  Can you get a product on the shelf in various stores? Is it brand strength/promise of marketing dollars for the brand?  Car dealerships work this way in the US.  Questions of persistence: how long will they last?  They do need some maintenance at least w/the presence of the product/service.  Perrier was off the market for 4 months for having traces of benzene; 3 years later, they still hadn’t recovered.  Careful about inferring long term impacts of brand effect.

Q: what’s the best way of dealing w/price?

A: it’s a variable that you control for. Price may be function of brand strength, but always control for it.

Q: it has to be a thought experiment though (the intercept point that’s # sales w/o marketing) b/c you can’t set things like distribution to zero (which was what happened with Perrier when it left the market).

Fit into legal dispute: in principle, many such disputes are possible. Any situation in which brand is monetized by someone who doesn’t own it: licensing, franchising, royalty rates/marketing support disputes, intracompany transfer pricing agreements w/tax consequences. Company argues that the brand valuation is lower than the tax commissioner thinks b/c of the marketing investment in local jurisdictions and so low royalty payments back into the US are justified.  Qs relate to duration of effects, how the brand itself is perceived/built in various markets.  Judges can be very sophisticated about how marketing works. Challenge: the model that courts use is grounded in tangible assets w/certain depreciation rates. You have to recast that in the light of marketing and branding, which are different.

My Q: how if at all do these standard approaches work for a small business that doesn’t trade in multiple countries?  A business that targets a subset of consumers, e.g., doctors?  Easier or harder to calculate?

A: There is no validated methodology for small business. You need the data of a publicly traded company. You can do something but can’t get an ultimate number for the value b/c experts have to make assumptions, which end up controlling the outcome (e.g., what’s the useful lifetime of a brand?). Even with the public companies you’re getting ranges like $600 million to $3 billion. For a small business: You can use something like a multiplier and find comparables that scale them up. It’s just not feasible with a sufficient degree of precision.  [It strikes me that this answer is super helpful to proponents of injunctive relief in cases of infringement that don’t involve global brands.  So I might aim this kind of expert at an irreparable harm finding.]

Q: why not ask what would happen if all other brands’ marketing went to zero too?

A: looking for a differential.

Q: sure, but you’d get a different answer. You need an answer to the Q of what an appropriate benchmark is, which is an issue with operational definitions of brand value.

A: yes, to do it, you inevitably rely on assumptions. 

Q: sure, and if I’m opposing you I argue that you taking all competitors down to zero is a better construct.

Anindya Ghose & Avigail Kifer, NYU Stern & Cornerstone Research, Search Engine Advertising, Trademark Building, and Consumer Intent

SEO advertising is competitive and lucrative. Search engines are two-sided markets.  TM bidding: using keyword ads for TMs you don’t own can in theory increase exposure of lesser known brands and benefit consumers using the TM as a generic term (e.g., Kleenex) [or using it as a shorthand because it’s top of mind even though they have no particular preference for that brand]. But TM bidding might also in theory prevent/delay consumers from reaching the TM owner’s site or create confusion.

Antitrust Qs: Does TM bidding contribute to consumer confusion? (It depends on what the resulting ad looks like). Does restricting TM bidding inhibit consumer price search? Does it prevent the most relevant ads from appearing in search results?  Would consumer search costs be affected by restricting TM bidding? If you can’t see ads from Costco, Wal-Mart, etc. when you search 1800 Contacts, then you may be getting less relevant ads. But if you have to scroll through a zillion ads, search costs may go up.

Lots of academic literature on related topics, including effects of competitor’s ad on your ad and tradeoff effects of your ad on your organic link. Type of device used can affect search costs and how you react to them. Mobile search implies something different about intent of search as well.

Presence of ads can increase total clicks and coversions: Example: Nike w/no ad at all gets 7.9 million clicks on its site; if Nike buys its keyword, then 5 million clicks go to the ad and 8.1 million clicks go to the organic result. Presence of ad changes consumer behavior.  Q: does it affect the competitors? A: It depends.  Where does it come from? Either from market expansion or from existing competitors.  Touches on differing intent of search queries.

Navigational searches: what is the search for? Specific product name: Samsung Galaxy S7 Edge—you may want different retailers but not competing products (though you might want those too).  TM—maybe different.  Look at click distribution, click through rate (conditional on ad rank), conversion rate (conditional on ad rank), bounce-back rates, dwell times—might give you a sense of whether consumers realized they didn’t want the result very quickly.

1800 Contracts: 1800 got big companies to agree not to advertise on 1800’s keywords and vice versa.  Did that violate antitrust law?  [Disclosure: I was FTC’s TM expert.  They were 1800’s search engine data people; we have not previously interacted.]  We had 10 years’ data for Google and 5 for Bing.  We had granular metrics.  Very few (less than 1%) of the ads had price information.  In response to Q: yes, consumers could click and compare prices.  People who don’t search “cheaper than 1800” are not looking for prices. [That is a weird assumption.]  You have to imagine a but-for world in which these agreements didn’t exist.  What would have been the results? You have to know Google’s algorithm/ranking.  Also, just b/c ads hypothetically would have shown up doesn’t mean consumers would’ve engaged with those ads.  Nontrivial, easy to rebut assumptions.  [Which convinced a pretty hostile ALJ and also the Commission, by the way!  Though this is definitely not my area, I will point out that there was some data about what happened when these agreements weren’t in effect with particular competitors; those competitors did better in search/conversion.]

Regulators are taking on digital ads: EC fined Google $1.7 billion for digital ad restrictions.  GDPR and its fundamental flaws: SMEs have nontrivial compliance costs w/display advertising.  Algorithmic/data divide is the new digital divide.  Is it the savvy who monitor how and why they’re getting advertised to, or the marginalized?  [I just saw a tweet suggesting that GDPR use in the US for Microsoft products is much higher than in the EU, since Microsoft adopted a global rule.] 

Fake news: do consumers perceive FB/Instagram ads as having been endorsed by the platform?  Must Instagram influencers disclose relationships when endorsing products? Can consumers differentiate b/t ads and other content? 2018 class action against Fyre media. This will be an important issue.  Distributional effects will be important. 

New tech by blockchain can help data quality.  Until 2017, we knew that the system had limitations: a lot of fraud.  $30 million every day for hackers arrested by FBI. Because of lack of transparency in digital ecosystem, there are disputes b/t authorities, platforms, intermediaries, ad agencies, consulting agencies like Accenture (which are now ad agencies), etc. Blockchain introduces much needed transparency in the system. Fraud and its magnitude would be hard to prove in past.  Now we have the data.

Q: price per click—is it long for this world? Paying for clicks is uncomfortable for the advertiser. 

A: we’ve gone the other way. Ability online for firms to identify, quantify path to purchase journeys has dramatically include. 

Q: why not bid for conversions?  [A: Some of it is.]  Why isn’t it all conversion?  [My related suggestion: Because marketing is still fundamentally probabalistic in a lot of ways?]

A: many products, people don’t buy on the internet. Consumers still want to go to your website, examine alternatives. Attribution issues are getting easier but still exist.  If you’re searching while you’re in the store, for example, we may be able to connect those. 

Q: is there an argument to be made about keyword ads causing dilution?  Even if the consumer isn’t harmed, the brand is.

A: hasn’t looked at it.  We had a branding expert.  [My answer: US law specifically allows comparative advertising as an exception to dilution; there is no legal support for the dilution argument here.]

No comments: