Personalized Pricing Improves Ecommerce Conversion Rates

Do you know the conversion rate for your online store?  You probably understand how many sales you make; that’s a pretty straightforward metric.  Understanding the conversion rate however, involves dividing the number of sales by the number of visitors or sessions.  A simple example is if you have 10 sales and 100 visitors, your conversion rate is 10%.

There are a few ways you can increase the number of users converting to sales in your e-commerce store.

Generate More Sales Online

In order to generate more online sales, it would seem logical that if you increased the number of visitors to your site you would increase conversions.  Keeping your 10% conversion rate, you would make twice as many sales (conversions) with twice as many visitors. The downside, however, is that this often requires running expensive ads; when the ads are done showing, the number of visitors drops back to previous levels. And there’s no guarantee the same percentage of users will convert.

How to Improve Conversion Rates

If you could bump your conversion rate up to 20%, with the same number of visitors, you would spend less to double your sales.  In order to do that, you can employ a variety of marketing tactics.  

User experience is very important - the fastest way to turn customers away and lose sales is to make your e-commerce store difficult to use and navigate. Strip the clutter, make it easy to find your products, and make it simple to check out.  Give them the information they need, some high-quality images, and a button to add it to their cart.

As your eCommerce site grows and converts more visitors, testing different methods against each other (referred to as A/B testing) will help you understand what is working and what is not.

Using Personalized Discounts Leads to More Profit

Incentives are a great way to move people from browsing to buying.  An incentive usually comes in the form of a discount.

Offering a site-wide discount where every visitor gets to use it, often isn’t very effective, especially in the long-term. Sitewide discounts simply train users to wait to buy until everything is on sale. Instead, using the Leaflet app, you can use the behavior data from your shoppers and offer personalized discounts to increase your profit.

See More Growth in Your eCommerce Store

Using the Shopify pricing app from Leaflet, you don’t need a lot of store traffic and you don’t need millions in sales.  Even with minimal data, the app will analyze the visitors to your eCommerce site and then, through sophisticated algorithms and AI, Leaflet offers personalized discounts to incentivize the sale. 

Increasing conversions and profits in your e-commerce store aren’t always easy, but with Leaflet you’ll get an edge on your competitors.

Prime for Day

Prime Day.jpg

We’ve just come off another year of Prime Day. The annual event has become a significant driver of success for Amazon. It was recently reported that more Americans are members of Amazon Prime than voted in the 2016 election or are members of an organized religion.

This raises an important question. What is so great about Amazon Prime that it has achieved a more widespread consensus than any other social construct in America? And how long can Amazon sustain that phenomenon? Eventually other market actors will catch on and offer programs with benefits comparable to Prime - right?

Shopify recently announced that they are investing USD $1 Billion to build logistics and fulfillment infrastructure to help their merchants compete more effectively with Amazon. Is this the beginning of a new day, where any merchant can create their own version of Prime? Or is there something more to Prime than just logistics?

Zero to $5 million in 24 months

My first startup was called AttorneyFee - it was an online marketplace for consumer legal services with a focus on price discovery and transparency.  I started the company with two cofounders in the Spring of 2012. Fast forward twenty four months to Spring of 2014 and we were generating revenue at a rate of $5 million per year.  At that point we began to attract attention from potential acquirers, and by August of 2014 we had completed the sale of our company to LegalZoom. All without a dollar of venture funding.

I’ve never told anyone the complete story of how I accomplished this - until now.  In this unprecedented tell all, I will explain:

  1. Exactly what I did on a step by step basis to grow revenue so quickly

  2. How we attracted acquisition offers, and

  3. The lessons I learned along the way

As a preface, and in the spirit of humility, I recognize that the learnings gleaned from this experience may not apply equally to other people’s situations.  Take what works for you and disregard the rest. I also recognize that every success is predicated, at some level, upon being at the right place at the right time.  Finally, I recognize that this the success I had with AttorneyFee was relatively limited. Having said that, there are many people who would appreciate having a few extra million dollars to their name.  I wrote this for them.


In order to appreciate this story it’s necessary to first explain a bit more about what AttorneyFee did.  AttorneyFee performed several functions for consumers: (a) it enabled them to quickly drill down on a comprehensive list of local attorneys who are experts in their issue, (b) it enabled them to find out how attorneys typically bill for similar cases, (c) visualize the range of average fees in their local market, and (d) schedule an initial consultation.  For law firms we offered the following functions: (i) attracting the attention of relevant consumers online, (ii) getting awkward conversations about the price of services out of the way, and (iii) connecting the consumers to the law firm either via a scheduled appointment or an inbound phone call. And that was when we would get paid, either when a consumer on our website booked an appointment with the law office or when they picked up the phone and called.

We didn’t always have that revenue model.  We started by selling long term advertising contracts that were not predicated on results - 12 month agreements.  We found that advertisers wanted more flexibility, so we reduced our contract lengths to 3 months, then to 1 month, and finally we moved from a monthly fee to a fee that was based on measurable results - appointments and inbound phone calls.  Our decision to modify the pricing model was not a sign of failure but rather a sign of success: we found that by reducing the level of commitment that we required of new customers we were able to get a higher percentage of prospects to give us a chance, and that difference led to a quantifiable change in the efficiency of our sales process.

But I’m getting ahead of myself.  The first thing we did is build out our website.  The hard part there was not the website itself, but collecting all of the data required to power the website.  Once we built the website and a solid prototype of the process for collecting, cleaning, aggregating, and analyzing the data and presenting the results of that analysis on the website in a way that was helpful to consumers, we then focused on growing organic traffic to the website via search engine optimization.  We were reasonably successful in this effort, quickly cresting over 100,000 unique visitors per month. Once we had the traffic, that’s when we started prototyping the business model, which meant cold calling law offices and trying to convince them to allocate part of their advertising budget to our company. circa 2012

The business plan worked well, for a certain amount of time.  We were able to start generating profitable revenues and grow those revenues quickly.  But eventually we ran into a few problems.

  1. Opaque value - We found that the no-show rate for appointments that were booked through us was problematically high, and a similarly a non-trivial share of the telephone calls that were coming from our site were from people who did not have a bona fide legal issue.  In other words, we found that even though we had shifted our revenue model to be as performance oriented as possible, the advertisers complained about the wide variation in quality of appointments and phone calls. In order to address these concerns, we dedicated a significant amount of product development resources to building systems and processes for qualifying the appointments/phone calls.  This enabled us to “take the high ground” within our industry in terms of establishing a reputation for quality and trustworthiness. It also enabled us to charge premium pricing relative to our competitors.

  2. Traffic plateaued - While we found it relatively easy to go from zero to 100K unique organic visitors per month, we found that going up another order of magnitude to one million unique visitors per month was a much more challenging task.  The organic traffic didn’t scale so easily beyond a certain point, so we started to supplement our revenue growth with advertising, which cut into our margins (more on that below).

  3. Problems of scale - as we grew we learned an important lesson about scaling.  Imagine a problem that is statistically unlikely to happen - let’s say there’s only a 1/10th of 1% of a chance of this problem actually occurring.  That’s a small enough chance that most entrepreneurs will say “I’ve got bigger fish to fry”. But if your business starts doing 10,000 transactions per day, then the problem that wasn’t worth your attention will now occur 10 times per day.  The point is that as you scale, statistically unlikely events start happening more and more frequently, so problems that you didn’t expect to encounter all of a sudden start materializing.


As mentioned above, when our organic traffic plateaued we started experimenting with pay-to-play channels to power our continued revenue growth.  We tried everything you can imagine - and some of the stuff you can’t.

    1. PR

    2. Search engine marketing (aka Google ads)

    3. Display advertising

    4. Radio Advertising

    5. Partners/Affiliates

The one thing we really didn’t get into was social media.  The reason for this is that most legal needs are “just in time” needs, meaning that you need to catch the customer at just the right time in their life for your services to be relevant to them.  We’ve all seen advertisements for DUI lawyers, but unless you were facing a DUI charge at the time, you probably never got out a pen and wrote down the lawyer’s phone number for future use. This feature of the consumer legal industry made search engines a more efficient venue for advertising than social platforms.

As explained above, we were able to establish premium pricing relative to our competitors. As it turned out, that proved to be an advantage in terms of acquiring traffic through paid channels.


Many people have asked me what it was like to go through an acquisition.  How does it work? How did the relationship get established? In our case the initial contact was made by my cofounder.  She was participating in a conference call hosted by the American Bar Association, and it happened to be that a high ranking officer from LegalZoom was also on that call.  Given the fact that our platform was about getting lawyers more business (whereas LegalZoom’s platform was more focused on replacing lawyers), we figured that the wisest thing to do was to publicly trash-talk LegalZoom.  In retrospect that was immature (even if effective).

That conference call led to a first meeting over lunch.  During that meeting, we ideated ways that our companies could partner together.  One of those ways was for AttorneyFee to provide LegalZoom with data about the average cost of certain legal services at the local level.  The idea was that they would use this data on their checkout pages (as a way of independently validating to consumers that they’re getting a good deal).  We developed this concept and actually operationalized it. This initial collaboration gave rise to a relationship that developed over time. As the relationship developed our two teams began to see more and more common opportunities, and eventually it just made sense to do a deal.

AttorneyFee was officially acquired by LegalZoom on August 8, 2014.  Our product was rebranded as LegalZoom Local, and all of our employees went on to great opportunities at a company with significant brand recognition.  I stayed on as the CEO of the business unit for a year, before I set off to pursue my next entrepreneurial journey.


  1. Sell something people want - we all know that the pace of innovation is accelerating.  This creates a temptation to build for the future. But the truth is that the best predictor of the future is the past.  Markets that were large in the past are, in most cases, going to be large in the future (even if they look and act different than they did in the past).

  2. Clarify your value proposition - sometimes the hardest thing to do as an entrepreneur is deciding what you’re not going to do; what your focus is going to be; and why the world needs your business to exist.  Don’t be afraid to revisit these questions down the line and update your answers.

  3. Learn to make advertising work for you - if every gas station owner needed to drill their own oil, they would never have time to run their business.  The point of the metaphor is this: traffic, like oil, is a commodity. If you figure out how the market for that commodity work, there is no need to create your own supply because you can purchase a virtually limitless quantity of the commodity on the open market.  

  4. You can succeed while doing dumb things - lots of people think you need to be brilliant to succeed in business.  You don’t. Relatively unintelligent people succeed in business on a regular basis, and relatively intelligent people succeed in business on a regular basis despite doing relatively unintelligent things.  Many businesses are able to succeed and thrive because they figured out one or two things.

  5. Remember to keep your cool - entrepreneurship is psychologically brutal.  Surround yourself with people who you can be honest with, and establish for yourself some regular routines that will promote your physical and mental health.


I hope you’ve enjoyed reading this story and that something in it resonated with you.  Stay tuned for more articles on the topic of entrepreneurship, revenue growth, and pricing.

- Richard Komaiko

Announcing Leaflet's Public API


We are excited to announce the upcoming release of our public API. Leaflet has already created a cutting edge artificial intelligence engine for personalized pricing. With the release of a public API, any developer will be able to easily incorporate intelligent, personalized pricing into their applications.

The planned documentation has been posted online. As of today, we are officially opening the waitlist to use the API. We will be releasing Leaflet to the waitlist in small batches. To save your spot in line, enter your email below.

This is a major milestone in our product roadmap. I’d like to express my gratitude to the Leaflet team, our excellent customers and partners, and all of our friends and family who supported us to get to where we are today.

Personalized Pricing and the Law

image 2.jpg

The legality of any practice will depend on what jurisdiction you’re in or what legal jurisdiction you are subject to. In the United States, there are laws pertaining to price discrimination at both the Federal and State level.

At the Federal level, the relevant law is called the Robinson-Patman Act. It would be beyond the scope of this article to provide a detailed analysis of the law in all fifty states. But for the sake of example, we will consider the laws of the state of California.


Federal Level - Robinson-Patman Act

A seller charging competing buyers different prices for the same "commodity" or discriminating in the provision of "allowances" — compensation for advertising and other services — may be violating the Robinson-Patman Act. This kind of price discrimination may give favored customers an edge in the market that has nothing to do with their superior efficiency. Price discriminations are generally lawful, particularly if they reflect the different costs of dealing with different buyers or are the result of a seller's attempts to meet a competitor's offering.

According to the Federal Trade Commission:

The Supreme Court has ruled that price discrimination claims under the Robinson-Patman Act should be evaluated consistent with broader antitrust policies.


In practice, Robinson-Patman claims must meet several specific legal tests:

  1. The Act applies to commodities, but not to services, and to purchases, but not to leases.

  2. The goods must be of "like grade and quality."

  3. There must be likely injury to competition (that is, a private plaintiff must also show actual harm to his or her business).

  4. Normally, the sales must be "in" interstate commerce (that is, the sale must be across a state line).

Competitive injury may occur in one of two ways. "Primary line" injury occurs when one manufacturer reduces its prices in a specific geographic market and causes injury to its competitors in the same market. For example, it may be illegal for a manufacturer to sell below cost in a local market over a sustained period. Businesses may also be concerned about "secondary line" violations, which occur when favored customers of a supplier are given a price advantage over competing customers. Here, the injury is at the buyer's level.

The necessary harm to competition at the buyer level can be inferred from the existence of significant price discrimination over time. Courts may be starting to limit this inference to situations in which either the buyer or the seller has market power, on the theory that, for example, lasting competitive harm is unlikely if alternative sources of supply are available.

The seller must inform all of its competing customers if any services or allowances are available. The seller must allow all types of competing customers to receive the services and allowances involved in a particular plan or provide some other reasonable means of participation for those who cannot use the basic plan. A more detailed discussion of these promotional issues can be found in the FTC's Fred Meyer Guides.


State Level - California’s Unfair Practices Act

California has its own statutes that govern price discrimination. These statutes are set forth in California’s Unfair Practices Act and are codified at California Business & Professions Code §§1740-45.


These statutes are similar, but not identical to the Robinson-Patman Act, and the California courts have enforced them by adopting some but not all of the above-summarized federal standards. Compare 15 U.S.C. §§13, 13a and 13b with California Business & Professions Code §§17040 et seqSee generally ABC International Traders, Inc. v. Matsushita Electric Corp., 14 Cal.4th 1247, 1256-64 (1997) (extended discussion of the history, purpose and reach of California’s Unfair Practices Act); and Harris v. Capitol Records Distributing Corp., 64 Cal.2d 454, 459, 413 P.2d 139, 142-43 (1966) (discussion of the differences between the Robinson-Patman Act and California’s prohibition of certain kinds of price discrimination under the Unfair Practices Act).

In general, the California state is concerned with business to business transactions and threats to the competitiveness of the marketplace. The exact prohibition is quoted below:

The secret payment or allowance of rebates, refunds, commissions, or unearned discounts, whether in the form of money or otherwise, or secretly extending to certain purchasers special services or privileges not extended to all purchasers purchasing upon like terms and conditions, to the injury of a competitor and where such payment or allowance tends to destroy competition, is unlawful.

California Business & Professions Code §17045.



The legality of personalized pricing will ultimately depend on what legal jurisdiction your business is subject to and the nature of your application. In general, American law on this subject is geared toward discriminatory B2B transactions and anti-trust concerns, rather than punishing retailers for B2C pricing that is designed to enhance the competitiveness of the marketplace. As with all legal questions, of course, you should consult a qualified expert to provide you with a reliable opinion.

Landmark Court Decision Impacts Businesses Everywhere


On November 27th, the 9th Circuit Court of Appeals announced its decision in a landmark case concerning the Telephone Consumer Protection Act. This case is now the law within the 9th Circuit, where many of America’s technology companies are located. This case clarified the definition of “automated telephone dialing system” (ATDS), as used by Congress. The question at issue was whether a system that sends text messages to a pre-defined list of customers is considered an ATDS, or whether the system must have a feature whereby it randomly generates numbers to receive those text messages in order to be considered an ATDS.

The practical consequence of this question is significant. If a system is considered an ATDS, then a business may not use it to send marketing text messages to a consumer unless the business has explicit written consent from that consumer. If a system is not considered an ATDS, then no consent is required. The penalty for sending marketing text messages via an ATDS without consent are steep, ranging from $500 to $1,500 per violation (100 messages to 200 recipients = 20,000 violations = $30,000,000 in penalties).

As of September 20th, every business that is sending messages from within the 9th Circuit and every business that is relying on a business located in the 9th Circuit to send text messages on its behalf should make certain that they have the required consent for the messages they are sending.

The top 5 pricing mistakes made by retailers


In our work helping retailers manage pricing we have noticed a few common patterns.  Some of those pattern behaviors are rational and efficient, while others are irrational and/or inefficient.  The following are the five most common pricing mistakes we see.

1. Not updating frequently enough

Some retailers set their prices a single time and then never reconsider them, or reconsider them too infrequently. The time requirement is definitely a deterrent. But if you’re making this mistake, you need to understand that the prices you charge are at least as important to the success of your business as anything else. Price is a huge driver of consumer demand. Price is often the factor that determines whether a visitor will convert to making a purchase or bounce without making a purchase. Price is half of what determines your margins. If you’re not paying attention to these factors and tending to them on a regular basis, it’s like you’re ignoring the instruments of success.

2. Updating without a plan

Some retailers approach pricing without a comprehensive plan.  The first step to creating a comprehensive plan is to form a solid understanding of what data assets are available for making decisions.  The next step is to define a strategy that leverages the available data, and make sure that strategy is understood by all relevant stakeholders.  Finally, develop a framework for incorporating your learnings.  This last step is absolutely indispensable.  Trying a little of this followed by trying a little of that is a waste of time, unless you know how to apply your learnings from trying this to your efforts at trying that.

3. Measuring the wrong way

Some retailers have a tendency to want to compare before and after.  This is called a "longitudinal study".  This is frequently not the ideal way to measure success for pricing experiments.  For one thing, there could be factors at play during one time period that are not at play during another time period.  As an example, imagine there was stormy weather in an earlier week and sunny weather in a later week.  It's almost certain that revenue from umbrella sales will be higher in the stormy week than in the sunny week, regardless of what pricing experiments were running during the respective time periods.  Because of this, the preferrable way to evaluate the impact of pricing experiments is to compare two groups from the same period of time: one group which will receive the normal experience and another group which will receive the experimental experience.

4. Ignoring canabilization

Some retailers use pricing as a mechanism to introduce their customers to new products.  They frequently will evaluate the successfulness of these measures based on whether or not the customer purchases the new product, or whether they continue to purchase that new product.  The problem is that purchasing that new product, in some cases, means not purchasing a product that otherwise would have been purchased.  In other words, a promotion of one product may come at the expense of some other product.  So in order to know whether or not the retailer is actually coming out ahead, a wholistic perspective must be adopted.

5. One size fits all pricing

Many retailers determine pricing on a one-size-fits-all basis.  Most of us know at an intuitive level that one-size-fits-all often really means one-size-fits-none.  People are not fungible.  Every person is unique, with their individual tastes, preferences, and budgets.  Ignoring these factors causes many retailers to experience suboptimal conversion rates and unnecessarily low margins.

divider 2.png

Leaflet’s AI enables retailers to automatically deliver true 1:1 price personalization without complex rules

The Cold Start Problem

isbil 2.jpg

Anyone from Chicago or New York is familiar with the problem of trying to start their automobile during the winter.  But that's not the type of "cold start" that this article is meant to address.  In computing, a "cold start problem" refers to a situation in which a computer system cannot draw any inferences for users or items about which it has not yet gathered sufficient information.

As an example, imagine a retailer that wants to use artificial intelligence to determine which product to recommend to their customers.  In the past, this particular retailer has only ever sold hats.  But now, the retailer sells hats and shoes.  The retailer may have plenty of data about how their customers respond when presented with offers for hats, but they will have zero historical data about how their customers respond when presented with offers for shoes.  This type of cold start problem is triggered by the introduction of a new item.

Another type of cold start problem is triggered by the introduction of a new customer.  The retailer may have tons of historical data about how its existing customers shop.  But when a new customer comes to the store for the first time, the retailer has no historical data about this person.  Yet another variation is the entrance of the retailer into a new community.

In all of these cases, the retailer cannot rely on historical data to make decisions.  The same dynamic applies in the realm of pricing.  Take a given product...if you've only ever sold it at once price point, you have no idea how sales would change if the price were increased by one dollar or decreased by one dollar.

Companies that provide pricing technology have taken a few different approaches to overcoming the cold start problem:

  • The Customer Approach - only work with retailers that already have a huge volume of historical data. Volume alone, however, is insufficient. In this approach, the retailer also needs to have a sufficient variety or diversity of historical data. The obvious disadvantage of this approach is that it pretty much excludes small to mid-sized retailers.

  • The Data Approach - only work with retailers that are willing to share or pool their data with other retailers. By combining data from multiple retailers, you can hack your way to a sufficiently voluminous and diverse variety of historical data. The catch here is that many retailers are reluctant to pool their data with other retailers - particularly with competitors - and with good reason.

Due to the drawbacks mentioned above, Leaflet has rejected both of these approaches.  Instead, our approach has been to develop algorithms that are not limited to repeating the past.  Our algorithms actually ideate and test new strategies, independent of historical data.  In this way, we can serve a wider variety of retailers without requiring that they pool their data with competitors, and without compromising performance.


The Complexity of Pricing Decisions

pricing001 2.jpg

Imagine a scenario where your store sells a certain product - we'll call it product A.  Perhaps there are other stores in your market that also sell product A.  Some of them will sell product A at a higher price than you, some will sell it at a lower price than you, and others will sell at the exact same price as you.

Let's imagine a scenario where 30% of competitors are priced higher than you for product A, 60% are priced lower than you, and 10% are priced equally with you.  Does this mean you should decrease the price you charge for product A?  Not necessarily.

  • Perhaps decreasing your price for product A will cause an increase in the velocity of sales of product A. But perhaps that increased velocity of sales is not due to additional customers or additional frequency, but rather, the effect of substitution as customers who would have otherwise purchased other products similar to product A now purchase product A instead. Determining whether or not this has occurred requires careful analysis of all the sales of all of the other products you carry that may be in competition with product A.

  • Perhaps the increase in velocity of sales is not due to substitution effect. Perhaps it represents a genuine change in shopping behaviors in which customers are devoting a larger share of their budget to you. But at what cost? If you reduce your margin on product A beyond a certain threshold, perhaps it was no longer profitable for you to have won that additional business.

  • Or perhaps the economics around product A are not the entire story. Perhaps even if you are not making money on product A, the transactions in which your customers purchased product A also included other products, and perhaps the average order value or margin profile of those products increased in conjunction with the decrease in price of product A.

  • Alternatively, perhaps decreasing your price for product A will not cause a measurable increase in velocity of sales of product A. This is a situation that economists would describe as price inelasticity. In such a case, it is possible that you just eliminated a profit center for your store for no reason.

  • All of the same possibilities exist in the inverse too. If you are priced below market on a certain product, it's not necessarily true that you should increase your price.

What further complicates all of this is that in the real world, a retailer's choices regarding the price of product A are not a binary (up or down), but rather, one of degrees (increase by $1, $2, etc).  And obviously, if we are dealing with a retailer that has multiple store locations, or multiple sales channels, the math is further compounded.  Finally, in the real world, the price of product A is not the only moving piece.  The prices of products B through Z are also in play, and there is bound to be a certain amount of interactivity.

In order to make informed decisions about pricing in competitive markets, it is essential that retailers approach the problem from 360 degrees.  This means taking into account every cause and effect relationship, no matter how large or small, and forming an understanding of those relationships as they aggregate with one another.  This is a challenging amount of analysis for anyone to balance in their head, or even in a spreadsheet.

divider 2.png

LEAFLET was created to help retailers move this analysis to the background by utilizing state of the art machine learning algorithms.  If you would like to learn more about how Leaflet can help your business automate and optimize its pricing strategy, please click below.

Fairness in Pricing

value price.jpg

Dynamic pricing strategies specifically designed to maximize retailer profits can nonetheless work to the advantage of consumers  

The following example illustrates how this can occur:
Joey likes Butterfingers, but Amy not so much.  Butterfingers normally sell for $2.00.  Joey’s indifference price - the price at which he’s indifferent between making the purchase or not - is $1.75.  Amy, due to her different set of tastes, wouldn’t be interested in owning the Butterfinger unless it was priced down to $1.00.  Now imagine the local candy store runs a sale for 50¢ off Butterfingers, or $1.50.  Since the sale price is higher than Amy’s indifference price, she will not redeem the deal.  Joey will redeem, but he would have been happy to pay $1.75, so the store effectively “overpaid” for Joey’s enjoyment of the candy.  The 25¢ spread could have been used to encourage Joey to add a complimentary product to his cart, try something new, or introduce him to a brand. 

Enforcement Ramp Up


Trade publication Marijuana Business Daily reports that "The California Bureau of Cannabis Control sent a cease-and-desist letter last month to Irvine-based Ghost Management Group, the owner of popular marijuana dispensary finder, ordering the company to stop advertising unlicensed operators or face unspecified “criminal and administrative penalties.”

On January 1, 2018, California's new Medicinal and Adult-Use Cannabis Regulation and Safety Act went into effect ("the Act").  Among other things, the Act created a new enforcement mechanism called the California Bureau of Cannabis Control.  Prior to the effective date of this Act, the distinction between "legal" businesses and "illegal" businesses was not so formal, and enforcement was mostly left to local police. 

Under the new regulatory environment:

  1. There is a formal distinction between "legal" and "illegal" businesses (holding a license is the ground floor for being considered a "legal" cannabis business); and
  2. Advertisements for cannabis businesses must clearly state the license number of the business that is responsible for the advertisement.

According to Bureau Chief, Lori Ajax, Weedmaps continued into 2018 to show advertisements from unlicensed businesses, and failed to include license numbers from any licensed businesses.  Old habits die hard. 

Compliance is Not Evenly Distributed


Recently there has been a fair amount of news regarding compliance problems within the regulated marijuana industry.  


In December 2017, police raided Sweet Leaf, one of the largest chain of retail stores in Denver.  At issue was the fact that the retailer was enabling customers to purchase amounts of product far in excess of statutorily mandated limits.  Between the immediate legal ramifications, the consequential damages, and the public fallout, the incident caused Sweet Leaf to shutter its doors.  


In January 2018, the Oregon Liquor Control Commission ran a sting operation.  They sent into adult-use marijuana shops undercover agents who were below the age of 21 to determine whether retailers would sell them product, or turn them away as required by the law.  Out of 86 shops, 16 sold product to the underaged undercover agents, or roughly 20%.

The agency issued citations to the violating stores, which also could face fines or temporary license suspensions.


Also in January 2018, the Springfield Republican newspaper obtained and published a set of records from the Massachusetts Department of Public Health.  The records showed that the Department has conducted 327 inspections since the state's medicinal marijuana program launched in 2015.  Among the violations found:

  • Plants were thrown in a dumpster instead of being ground up
  • Inspectors were able to stores without being carded
  • Lack of security cameras, defunct generators, faulty alarm systems, and improperly labeled plants.
  • Inspectors also tested products and found contaminants that led to recalls.

Promotions Gone Wrong


Discounting too much can lead to economics that are not beneficial for retailers.  But the conventional thinking is that - at least from the perspective of customers - the more deeply a product is discounted, the more successful the promotion will be.  But what happened in France yesterday shows that this thinking is flawed.

As reported by CNN and many other major media outlets, a chain of supermarkets called Intermarché decided to run a promotion on Nutella.  The popular chocolate-hazelnut spread was discounted by 70%, and the result was utter pandemonium.  An unanticipated number of consumers flooded the stores, creating lines that lasted for hours.  People were grabbing, trampling, and fighting to get Nutella. 

"People just rushed in, shoving everyone, breaking things. It was like an orgy," another employee in Forbach, northeast France, told AFP. "We were on the verge of calling the police."  Another location was actually forced to call the police, as disturbances broke out.

There are limits to everything.  What this story shows is that there is a limit to how low a promotion should go.  It also serves to underscore that before launching a promotion, retailers must consider things like inventory at each location and elasticity of demand.

Promotions 101

Imagine your store wants to run a promotion on vapor cartridges...


1/2 Off

Next Wednesday only

Let's say your store has the ability to target this promotion to any subset of your customers based on demographics and past purchases.  That is to say, you have the ability to include or exclude specific customers from participating in the promotion.

Imagine there was a group of women who have demonstrated an affinity toward vapor cartridges by regularly spending at least $100 per month on this product category.  The question is, should you include them in the promotion or not?

The argument in favor of including these women in the promotion is that they like vapor cartridges, so they are likely to redeem the promotion.  The argument in favor of excluding them from the promotion is that they were likely to buy vapor cartridges anyway, so it's really more of a "discount" than a promotion.

So what's the answer?

The answer is that there is no one size fits all answer.  The only scientific way to even approach the question is through trial and error.

  1. Randomly divide the women into two groups;

  2. Make the promotion available to half of them but not available to the other half;

  3. Wait some period of time; and

  4. See which group ends of spending more money down the road;

The caveat is that changes in spending will not be evenly distributed within the groups.  That is to say, some of the women who received the promotion will go on to increase in spending, while others will decrease in spending, and others will be constant.

In an ideal world, someone from your team would roll up their sleeves and spend some time analyzing the data.  Are there any similarities between the women who went on to increase in spending?  What do they have in common?  And how could we know, for future promotions, who to include and who to exclude?

In the real world, this process is extremely time consuming, and there may not be anyone on your team with the bandwidth and acumen required to perform the analysis.  Furthermore, any lessons that can be learned from this exercise will be limited in scope.  That is to say, whatever insight you're able to derive about women who like vape cartridges will not help you when it comes time to market to men who like edibles.  Finally, its extremely uncommon for customers to only ever be exposed to a single promotion.  So measuring the effectiveness of a given promotion normally will require you to first develop the competency to distinguish the effects/influences of one from promotion from the effects/influences of another promotion.

Clearly it's not realistic to do this much math by hand.  But the truth is, even with the aid of a computer, without the right framework it will be an exercise in frustration.


Leaflet takes the complexity out of this process.  Leaflet is a first of a kind promotional intelligence platform.  Leaflet continually looks across time to identify the critical success factors that result in high value customer relationships, and automatically applies what it learns to create promotions that are personally optimized to each individual customer. 

For More Information