“A social platform with voice recognition would be valuable, as long as it respected privacy”

Reza Sabernia - edit
Reza Sabernia

Interview with Reza Sabernia, founder and CEO of BrainMustard.

Hi Reza, please tell us which services BrainMustard provide?

BrainMustard offers a new way to scan and analyze Internet chatter and social media. We build comprehensive models of consumer behavior within the brand ecosystem. These models help our clients to find revealing, and often unexpected, insights about consumers, which in turn help companies enhance the customer experience and increase sales, sometimes by an order of magnitude.

Which type of companies benefit from your services?

The companies that benefit from our services are concerned about the consumer experience. Currently, our client roster includes a bank, a commuter railroad, a pharmaceutical company, manufacturers, retail stores (including the largest sports retailer in Canada, with 1200 stores and an apparel company in India), as well as world-famous drink brand Diageo, who sells drinks such as Captain Morgan, Smirnoff and Guinness.

What is your background and what made you start BrainMustard?

I studied computer science, specializing in natural language processing and artificial intelligence, and then received my bachelor’s of commerce degree from the University of Toronto. I received my MBA from Kellogg School of Management at Northwestern University.

When I graduated, I worked in an employee recognition startup for over a year; however, I was working on the idea of building BrainMustard throughout that time. Previously, I worked in a firm that did social media security, trying to find Internet criminals online. The technology for semantic algorithms for understanding the Internet chatter was not that different from what we use for BrainMustard, so I learned about information flows online and how to analyze the content to see what constitutes valuable signals and what is noise. That knowledge helped me in forming the technology that is used for BrainMustard.

I saw that the tools used for social media monitoring and sentiment analysis were limited to tactical purposes. They’re tactical because they helped social media specialists and members of marketing departments read verbatim sentiments and itemized information, as well as engage in a direct conversation with users online, But there were no tools in the market that offered a big picture solution that professionals with more strategic roles could utilize to make decisions.

I also noticed that social media monitoring solutions specialists knew that end-users can only process a couple of thousand messages at best in any given month, while they had access to tens of millions of relevant content volume. Because they wanted 100% accuracy and relevancy, they threw away a lot of content. Some of that content was irrelevant, but a substantial amount of the discards embodied interesting information. Because of this practice, brands didn’t get all of that information, and were losing important data.

Based on that need in the industry, I created BrainMustard. We don’t throw away any messages; instead, we allow them to get crystallized into thematic clusters and organically form modules that can be studied. These modules can offer new and fascinating insights from social media. We have a bottom up approach; if cluster is irrelevant, it will be deleted in the very end when we know what it means, but it will not be thrown out just because we want 100% accuracy. We don’t mind examining a lot of noise in order to capture all the information that’s out there.

What are your responsibilities in your current role at BrainMustard?

I am the Founder and CEO of BrainMustard. I supervise technology and business development, and meet with new clients to see what their needs are and if we can offer value to them.

A big focus of mine is on innovation. For example, we are building technology to track customer behavior inside stores. We are working on identifying new ways to flag and tabulate information, which is a never-ending process. We work closely with our clients to see how we can offer them meaningful information. A big portion of our innovation comes from solving the problems for our clients.

What is the current focus for BrainMustard in the near future?

Our current goal is to offer generalized industry reports. We have been working with brands directly, but are now trying to create syndicates that would be useful for all the brands in a particular industry. It’s a challenge so far, but we hope to have it ready in the next few months. Currently, our focus is on customized customer experience maps and social influence maps, which I would say is the core of our business and an area where we outperform our competitors.

Can you give specific examples where one or more of your clients have made changes in their communication, products or similar, based on the information or analysis you provided?

Bauer Hockey came to us because they are dealing with a changing marketplace. To increase hockey popularity among millennials and immigrants, Bauer Hockey collaborated with NHL to build Wal-Mart-size flagship stores in the original six cities where the NHL started. Those six cities are really the hockey mother-ship, the places where hockey is the most popular. The goal of the stores is to offer a memorable hockey experience that resonates with the parents and the kids. Bauer wanted to know what matters most to parents whose kids play hockey. They believed that a major concern for parents was safety, and they wanted to make the focus of the stores safety.

However, after BrainMustard provided an analysis of the hockey eco system, it became evident that while safety was an issue for parents in general, it was not the main issue for parents of children who actually played hockey; instead, what was concerning to them was the ice time their kids would get during weekly practices and games. When the first store opened in Boston, its main focus was on the improvement of performance in order to allow children getting more ice time. Because of the innovative information that BrainMustard provided, the store is a huge success.

Another example is of a startup that became a very successful franchise and was actually acquired by a major coffee and beverage company. The founder wanted to offer exceptional experience for tea drinkers, but the problem was that he was trying to do it in San Francisco, which is a saturated market. BrainMustard looked at the social spectrum and identified segments, such as ritualistic drinkers and the health conscious.

We also focused on segments that were overlooked, such as stay-at-home moms. These are women who are career oriented and successful, but who choose to stay home after having children. The problem is these women can feel disheartened when seeing their peers climbing the work ladder, while they continue to be just mothers. These moms not only feel like they are missing something, but they also don’t have much to share at social gatherings. They can only talk about kids, while other women talk about their jobs.

To tap into this market, the founder and BrainMustard came up with idea of creating gears for the tea drinking experience that is exotic and has a story behind it. The gears consisted of pots and saucers and other items. The items encourage ritualistic behavior and a certain level of preparation. The founder would sell gears and offer workshops to prepare exotic teas, which had a story behind them. This strategy was a huge success because moms could invite friends over and have a story to share about that experience.

What information from external sources do you use today to make your analysis, and which are the most important?

We use information from many companies; one of the better ones is actually Twingly. We use Twingly for blog sources; their service is great. When we have questions or requests, they responded quickly. We also get information about Facebook users from another source, and are provided with forum content, as well as having our own in house solutions. And there are many others.

Is there any consumer data that is difficult to retrieve today that would help you provide an even better service?

It’s all about privacy today; for instance, Facebook data is great, but most providers provide it as an aggregate form. While that is good for dashboards, it is not very beneficial for insight to know what words people use and what are the main associations between brands.

Another source that would help us provide better service would be comments from Amazon, which can offer value because they’re words from exact customers who have used the product and are sharing their experience. Pretty much the rest is accessible, which is good news.

Are there any social platforms that are closed today that you would be interested in tapping into for monitoring that would benefit your customers?

Facebook information would be very interesting. As well, millennials are active in teleconferencing and voice software, so if there was a social platform with voice recognition, that would be valuable, as long as it respected privacy.

What are your greatest challenges ahead at BrainMustard?

Our greatest challenges are in developing ways to communicate findings effectively. Our findings are numerous and we work hard to make the narratives behind them understood. We are developing new technologies that will correlate what people do either consciously or subconsciously in the stores with what they say on social media. We’re working on being able to segment the market based on people’s actual store behavior so our insights won’t be based simply on what people say on social media.

How do you think the business you are in will change in the next 5-10 years?

We are not in a business that is as fast moving as some pure technology players in the market. In terms of insights, that will not change much because societies don’t change as fast as technology does. The output of our solutions wouldn’t change as much as the backbone of the system, which will. We have to keep up with platforms to offer solid coverage; while the output will not change that much; the technology used for the sake of presentation will change. For example, augmented reality and 3D printers are new innovative ways to offer a more tangible and interesting output with additional dimensions to clients.

By Renata Ilitsky

Same same but different

My name is Björn Milton and I co-founded Twingly 10 years ago together with 3 other guys. For 4 years I worked as the CTO there. Now I am the CEO of another startup called Roombler.

I still work with Twingly as the chairman of the board. So I still put a lot of thought into the company. When I’m not thinking about Twingly, the rest of my time is spent thinking about Roombler. So I have a pretty good picture about both of the companies. So I thought I’d write a post about some of the similarities and some of the differences that I see between the two companies.


As you probably know if you’re a frequent reader of this blog, Twingly deals in data. Blog data from around the world is gathered, processed and analyzed. This data is then distributed to customers, mainly via two different products. It is very much a question of making sure that we discover and collect as much data as possible as fast as possible. Customers interacts with Twingly through APIs. They receive data on a continuous basis. The data is very much the product. It is computers, not humans, that interact with Twingly for the most part. Twingly has a global customer base.

Roombler – suitable for smaller hotels


Roombler is a mobile control panel for small accommodation services. Our customers include smaller hotels, vacation rentals, B&Bs, apartment rentals (think AirBnb) and so on. We make it easy for the operators and staff of these businesses to handle reservations/bookings, guests, prices and availability. We make sure that external systems (such as sales channels) are kept in sync. Roombler is a mobile app and is always available to the users. Roomblers customers are very global in nature and we already have customers on all the continents.


There are a lot of similar aspects between the two companies. I’ll go over a handful of them here.


Both Twingly and Roombler are B2B-companies. Our customers are other businesses that wants to buy our services. The opposite is a consumer driven model, B2C. In general I personally tend to like B2B better. To me it is a more concrete model where it is obvious that if Company A can provide value to Company B then Company B is willing to pay for that. It is also often much easier to pinpoint and reach out to the correct segments in the B2B case. In late years, we have seen an upswing of the B2B tech companies with high valuations and some IPOs.

Product based

Both are product based companies, we build one product that many will buy. The opposite would be a company that sells professional services. They sell their time and build custom solutions for each customer. I tend to gravitate towards product based companies since I like to realize my ideas, not the ideas of somebody else. Of course we listen a lot to our customers, but it is always from a product perspective. We don’t build stuff that we don’t think would benefit the product as a whole.


Both are SaaS based companies. That means that the products are sold as services where the customer buys the outcome of the service. In Twingly’s case they buy data and doesn’t care about how that data has been collected and stored. In Roombler’s case the customer buys the right to access Roombler service through an app on their mobile device. They don’t need to keep a server of their own to store the data, they can access the product from anywhere and they don’t have to worry about uptimes or updating their software.


A very common business model in connection to SaaS is the idea that the customer will pay over time. Instead of paying a large sum upfront, the cost is spread out over time to better match the cost with the value. In Twinglys case this is very natural due to the nature of the business. As a customer you’ll buy access to data that is delivered to you over time. It would be very hard to find an upfront model that would work in this case.

In Roombler’s case the property management system business (PMS) is very much coming from an old pay upfront type of model. This is not very ideal for the customer, especially not if you’re a small business. So, for our segment, and in connection to the SaaS model it is very natural for us to use a subscription based model.

Technology driven

Both companies are at the core driven by technology and big resources are spent towards making sure the systems are functioning and towards making continuous improvements. This means that we have more developers than sales people. It means that we have the ability to quickly build new things, to constantly innovate. Both companies are founded by tech people, by developers.

During the years I have seen a lot of startups that have been founded without having the tech people at the core, in the founding team. This is, in my experience, always a bad thing because it tend to lead to a situation where the technical parts are seen as some one off that can be purchased and then put to work. If you build a technology driven business (which most businesses are today) you need to have tech people and developers at the core of the business.


There are also a lot of differences between the two companies:


For Twingly, the customer is often a larger organization with departments and multiple levels of management. Mostly the actual person that does the deal is a business person of some sort. This person looks at it from a business perspective and doesn’t care that much about the actual interaction with Twingly’s systems. So at the 10 000 feet level, the customer only cares about the data. But there are of course more actors in the process that needs to be tended to, not at least the developers that will be doing the integration part.

Roombler sells to smaller, owner driven organizations, where the owner is most often the one we talk directly to. The product will also be used by the owner and therefore cares a great deal about the actual user experience. Roombler is the central business system of these organizations. If they cannot access Roombler or if we do something wrong it will greatly affect our customers. Roombler is mission critical to these businesses.


Twingly operates on a very well defined niche market. The number of potential customers are counted in the tens of thousands. It is a market that is largely underserved and our customers is constantly looking for better and more data. The market can relatively easy be processed with outbound activities. It is feasible to call or email every known potential customer.

Roombler’s market is much bigger in terms of potential customers. The potential customers are counted in the millions. This makes for an interesting challenge when trying to communicate with them. It is simply impossible for us to reach all of those customers by outbound actions. We very much need to build a inbound based process where the majority of the customers find us.


Twingly sells its products to relatively few customers. Each deal is worth quite some money and the sales process is often of some considerable length. The product is often sold to companies that in their turn takes the data and sells it to their customers. As such, it is easy to see that the product is adding directly to the customers revenue.

Roombler is much more a mass market product and we’re catering smaller businesses. As such the pricing point is much lower. As a comparison, Roomblers entry level pricing point is about 2% of Twinglys entry level pricing point. Also, Roombler is not as direct in adding revenue as Twingly.

So, there you have it. Some of the differences and similarities of the two companies. Each business is unique of course, but I think it is valuable to compare your business to others. That will help you understand your own business on a much deeper level and enables you to put some context around who you are and what you’re doing.

By Björn Milton