Other Cool Stuff

Princeton University Grad Writes Thesis About Pear Analytics

By January 24, 2010 December 3rd, 2012 3 Comments

The following editorial is actually a thesis written by Josh Lavine, a student at Princeton University whose task was to interview a start-up company, preferably in the hi-tech area for an Entrepreneurship class.  It is quite long, but describes how Pear Analytics was started and where we are going, the challenges we face, and more.  I’ve also left out the Appendix due to length, which you may see notated throughout the report.  This is Josh’s final thesis, and frankly I was impressed by how much he learned about our business and industry in the mere 4 or 5 hours he interviewed me.  I also plan to implement several of his suggestions for improvement which he notes at the end of his report.  Josh is personally invited by me to come join our team at Pear any time. Enjoy!

It is December 23, 2009, noontime. I smooth my shirt and try to lick the tomato sauce stain off my sleeve, then open the door to the office of Pear Analytics. What I saw surprised me.

The office, located in San Antonio, TX is quaint, but strangely chic. It is only one big room with no walled-in spaces, except for the two small conference rooms in back. Large neon green and blue balls are rolling around the floor (I would later find out that instead of buying expensive chairs, the team realized they could just sit on cheap, cool colored exercise balls). Ryan Kelly, founder and CEO of Pear Analytics is presently standing (towering, really—he’s must be 6’6”) at one end of the room, watching one of his office mates from BrandStack, the company he shares office space with, play tennis on the Nintendo Wii on an enormous flat-screen TV—probably the most expensive piece of equipment in the room. If Kelly turned his gaze down and to the right, he would see Romy Misra, his senior analyst, writing equations in dry erase marker on a large glass table. She records her calculations on her laptop. Kelly’s other employees—just three web developers—were out for the day. I follow Kelly into one of the back conference rooms. We sit down and start talking.

A Man with Many Interests

 

It is about 18 months since Kelly founded Pear Analytics, now a Web-based company that helps websites with Search Engine Optimization (SEO), and things are going well. But a decade ago he probably never thought he’d be doing something like this. In 1998, a recent graduate from the University of Massachusetts, Amherst with a BSME in Mechanical Engineering, Kelly got a job in the aerospace industry working for Pratt and Whitney, designing and managing jet engine production for the U.S. Air Force. During his first year there, he developed a new kind of airliner turbine foil, for which he received a propriety design patent in 2002. At Pratt and Whitney, Kelly was frustrated by the non-linear way engines were produced on the floor. He described the production floor as vast and organized by an arbitrary positioning of part manufacturing booths (so to speak). Kelly observed that oftentimes a part or an engine in production would take days to get from one stage of production to the next because the booths were on opposite sides of the room, and transportation required trucks that had to be reserved in advance. As an engineer he had little influence in getting the floor rearranged for smoother production. Wanting to become more involved in management, in 2001 Kelly entered business school at the University of Hartford, where he studied Marketing Management. In 2002, Kelly transferred to the University of Phoenix, where he finished his MBA in 2006. For the last year of his 5 year stint at Pratt and Whitney, he acted as the sales and marketing director of Pratt and Whitney’s $20 million spare part business in San Antonio.

In 2003, Kelly moved from Pratt and Whitney to Blue Clover, a creative design firm in San Antonio, where he served as general manager and partner. At Blue Clover, Kelly managed all business processes and the company’s investments, and oversaw all of the production staff. He also oversaw all financials, budgets, legal contracts, and company proposals, provided key web strategies, SEO and marketing techniques for many of the company’s clients.[1] With Kelly there, Blue Clover became distinguished as an award winning design firm. In 2006 it was honored with 16 Addy awards, including 5 gold for elements of advertising, interactive media, and collateral material; and in February 2008 it was named one of San Antonio’s best places to work.

At Blue Clover, while Kelly helped his clients build their websites and construct an image for themselves, he kept receiving the question, “Why is my website not number 1 on Google?” Blue Clover offered some SEO services, which Kelly then believed to be extremely important. An April 2006 San Antonio Business Journal article quotes him:

General Manager Ryan Kelly adds that Blue Clover offers tools like search-engine marketing campaigns that optimize site awareness. “It’s a key component,” Kelly says. “It’s like having a store. You can’t get there if you can’t find it.” He stresses that awareness is key, no matter how attractive or high-tech the site is. “We help you utilize your Web site, like an employee. We make it work for you.”

Kelly started becoming very interested in SEO around this time. A man with passionate interests and a mind to serve his clients as thoroughly as possible by being as knowledgeable as possible, Kelly started researching all there was to know about search engines, and he became something of an expert. In early 2008, Kelly sold out of Blue Clover, and in March of that year, he started his own company that would specialize in SEO consulting—Pear Analytics.

Brief Explanation of Search Engines

The first link that comes up in a Google search gets 42% of the clicks for that search term. The second gets 18%. The third gets 12%. All the rest of the links on the first page of the search get under 4%. The links on the second page all receive under 2%, and people rarely ever click on links on the third page and beyond.

This is because Google’s goal, and the goal of any other search engine, is to deliver the most relevant results to a given search term as quickly as possible. In order to do this, Google has developed a complex and completely secret algorithm which determines the relevance of every single website in existence to every possible search term. The algorithm is Google’s most highly guarded secret (ask anyone who works for Google about it and they will politely tell you they cannot talk about it at the risk of losing their job), and it is constantly changing. The algorithm looks at a number of factors when analyzing a website for relevance to a search term. A few of the most important factors I describe below.

Keywords. A keyword is a word that is relevant to your site. Most websites have, in the coding, a “keyword tag” where one can list words that one thinks are relevant to the websites. Keywords appear in the coding of your site in other sections too, as well as in the text of the site itself. There are a number of schools of thought on how to optimize keywords in a website, but Google takes into account a few things. Lots of people like to cheat with keywords by doing things like putting a whole bunch of keywords as white text on a white background in their website. When you load the site, you can’t see them—the background just looks white, but the keyword is everywhere in the text of the code. Google caught onto this practice and now discounts a website when the keyword is overpresent. People also try to cheat by including high volume search terms in their keyword tag that have no relevance to their website (like a shoe store putting “iPhone” in it’s website code). Google is capable of detecting this as well. But even if one manages to get away with it, a person who types “iPhone” into Google is highly unlikely to click on a link to a shoe store even if the link pops up because that’s not what that person was searching for.

In-bound links. In short, as Ryan Kelly put it, “search engines follow people.” Google does this in a complex way. It assigns a number between 1 and 10 to each website, which is that site’s PageRank. The algorithm for assigning this number is secret. Google describes what PageRank means:

PageRank relies on the uniquely democratic nature of the web by using its vast link structure as an indicator of an individual page’s value. In essence, Google interprets a link from page A to page B as a vote, by page A, for page B. But, Google looks at more than the sheer volume of votes, or links a page receives; it also analyzes the page that casts the vote. Votes cast by pages that are themselves “important” weigh more heavily and help to make other pages “important”.[2]

Basically, the more popular or important or linked-to the website, the higher the PageRank. Websites like Facebook, Twitter, Yahoo, MSN, AOL, and Google itself all have extremely high PageRank. Websites like the Princeton homepage have lower, and personal profiles on Facebook likely have even lower. The more links from other high-ranking websites that point to a given website, the higher that website’s PageRank will be, and the more likely that website will show up in a search pertaining to that website’s keywords.

Website structure. Google is capable of determining whether a website’s general structure is good or bad. If a website is poorly structured from a technical standpoint (relating to the site’s content, or to the sitemap, or to any number of structural aspects), it won’t rank high in Google.

Loading Speed. If a website takes too long to load, people lose interest and move on to something different. A test done at Google in 2006 showed that going from 10 to 30 results per page increased load time by about 0.5 seconds, but resulted in about a 20% drop in traffic. And tests done at Amazon in 2007 showed that for every 100 millisecond increase in load time, sales would decrease about 1%. Kelly and his team researched into this and wrote a short white paper called “How Load Time Relates to Visitor Loss” in August of 2009.  They borrowed data from the Google and Amazon studies, as well as from a study done by Akamai in 2006 which found that if your website takes longer than 4 seconds you lose one third of your visitors, and, using the model of a first order radioactive decay equation calculated an approximate graph plotting visitor loss vs. page load time. The graph, as well as the assumptions used to calculate the curve, are in Appendix 1.

SEO consultants try to determine what Google’s algorithm is, or at least, what aspects of a website it weighs as more important than others. Every SEO expert believes something different from the next one about what is most important, and many of them get good results using completely different methodologies. Kelly has his own biases and beliefs regarding Google’s algorithm, and he advises his clients based on them with measurable success.

The Evolution of Pear Analytics

In early 2008, Kelly’s consulting business was going well. He had a number of customers who were willing to pay him about $3,000 dollars/month for six months for his personal SEO services. After a few months, however, Kelly realized that he was doing exactly the same thing with each customer. He would gather data about their website—keywords, number and quality of inbound links, sitemap structure, webpage layout, load time, etc. and then diagnose the problems and develop a plan to fix them. The problem, as with every SEO consulting business, was that gathering data usually took an incredibly long time. The time was spent passively diagnosing problems instead of actively fixing them. But because the process was essentially the same, Kelly’s engineering background kicked in and he started asking himself if the process could be automated. He opened a blank PowerPoint document and wrote a series of mathematical if-then statements that would later become the algorithm for his automated website data gathering program called SiteJuice. His algorithm would tell a computer to “crawl” a website (scan it), and by gathering and analyzing a vast amount of data that would normally take a human days to collect, within minutes pop out a score, from 1 to 100 that indicated how optimal the website was. It would also provide an analytic breakdown of the website in many different SEO related categories: keywords, website structure, tagline, etc. Kelly describes the algorithm as “a series of nested weighted averages”, which means that it analyzes many separate factors, scores them based on his rubric, and then weights each of them in a final average to give one complete score. The weights Kelly determined are based on what he believes is important for SEO. (As aforementioned, there is speculation and debate among SEO experts as to what is most important. Kelly’s own beliefs bias the weighted averages in his algorithm.) The final breakdown assesses the site for the biggest weaknesses and helps Kelly determine what needs to be done to fix the website.

Kelly knew that he had written something original and potentially market disrupting, and so he needed to protect the algorithm. He decided that applying for a patent was not a good use of his time or money, so he opted for secrecy. But Kelly needed to convert his mathematical equations into computer code, and Kelly had little experience in writing computer code. He needed to bring someone on to help him, and he needed someone good that he could trust. In early 2008, a friend of Kelly’s introduced him to Vidyut Luther, an expert in system design and architecture who was doing some consulting of his own—at the time he was developing and implementing a mock trading game for J.P Morgan on Facebook. Kelly liked Luther’s attitude and trusted his capabilities, so Kelly hired Luther and they started developing the product in May of 2008.

Kelly’s vision for the product was one of utter simplicity. He wanted to take a very complex process and make it “stupidly user friendly”. He wanted to take the complicated results from his algorithm and explain them in, as he calls it, “Fischer Price language”, so that anyone, no matter their comfort level or expertise with SEO, could easily read, understand, and use the product. Kelly designed the website interface using Powerpoint and Apple’s slideshow presentation software, Keynote, then asked a friend to make it pretty in Adobe Photoshop.

Once the website was up and the product nearly developed, Kelly and Luther approached an SEO agency (name unknown, I will refer to it as “Agency”) in the 3rd quarter of 2008. By this time Kelly and Luther had written the code out and developed a website analyzer that they knew would work. Agency was interested in making a deal with them, and they planned the whole thing out. Agency would fund the further development of Kelly’s product, and in turn would have co-ownership of it. Things were looking good for Kelly until in the 4th quarter of 2008, the economy took its biggest downturn yet. Agency was forced to cut third party help because of a constrained budget, and they backed out of the deal with Pear Analytics.

These were dark times for Kelly. Kelly remembers not invoicing a single client for a period of about three months at the beginning of 2009. His consulting business was not providing enough money to bootstrap his product along, and he needed some investors badly before his product stagnated and he lost all of his customers.

In March of 2009, Kelly launched his website analyzer tool and entered a small IT company competition in San Antonio called Innotech Beta Summit. Kelly was shocked that he won first place, for as he said, the competition was fierce. Blair Garrou, managing director for DFJ Mercury, on of the venture capital investors present at the Summit confirmed Ryan’s hunch that people would gravitate towards “Fischer Price Language” in his explanation for why they chose Pear Analytics as the winner:

“The reason we chose Pear Analytics as the winner was because of how they took a seemingly technical and complicated process of analyzing the searchability of a Web site and put it into a non-technical, easy to understand tutorial. Because of that, the market for this application has immense potential.”

However, despite all the hype around Pear Analytics, the Innotech Beta Summit did not do much for Kelly other than give him exposure. The win was a nice pat on the back, but Kelly needed money and none of the investors there were biting.

Kelly entered another competition hosted by the Capital Factory Tech Incubator Program in Austin, TX. The winner of that competition would receive a cash investment of up to $20,000, more than $20,000 in free services, and mentorship from some of the top entrepreneurs in central Texas. Unfortunately, although Kelly did well (placing in the top 10 out of 250+ applicants), he did not win this competition. But he would later say that that was a blessing, for if he had won he would have had to move to Austin and he never would have met his incipient office mates at BrandStack.

Around this time BrandStack, a small startup that was creating an online marketplace for logo designers, was looking for cheap office space to rent, at the urging of their two primary private investors. The two investors were Morris Miller and Pat Condon, who had both been co-founders and key players in the development of Rackspace Managed Hosting (NYSE: RAX): Condon had been instrumental in developing Rackspace’ unique customer service mantra “Fanatical Support”; and Miller had acted as President, Co-CEO and Co-Chairman, and he had helped build the company and its subsidiaries including ServerBeach. Serendipitously, BrandStack and Pear Analytics began to share an office together. The guys at BrandStack liked Pear Analytics so much that very soon they put Kelly in touch with Miller and Condon.

Before talking with Miller and Condon, Kelly had been in touch with a potential investor who had expressed interest in Pear Analytics after the Capital Factory competition. However, this investor had been hot and cold with Kelly—being very interested one day and seeming disinterested the next. A trusted friend of his, Paul Singh, Pear Analytics’ “Business Intelligence Expert (aka ‘Mr. Metrics’, as he likes to be called)”, advised him not to have more than three meetings with a potential investor, because as he said, a good investor can tell you yes, but can also tell you no with equal conviction. This particular investor had been stringing Kelly along and meeting with him time after time was getting tiresome, but before he came in touch with Miller and Condon, this investor was his only prospect because, as Kelly put it, “I don’t run around in investor circles. I’m trying to develop my product.” Kelly was despondent. Pear Analytics was stagnating, and Kelly could not bootstrap his business for much longer without losing steam. He needed to accelerate his product development and get something out there for customers to start using. But he couldn’t do that without money. So without the time to explore other options he still pursued the hot/cold prospect, but with no answer.

The tides turned for Kelly in when, after three meetings Condon and Miller, they decided to invest $100,000 each in return for an undisclosed amount of equity in the company. Miller explained his reasons for investing in Pear Analytics. As a co-founder of Rackspace, Miller was very familiar with SEO and website development. He said that before he met Kelly, friends had recently ask him on five or six separate occasions, “Why is my website not number one on Google?” Miller’s philosophy is that good business happens in “markets where people need stuff.” In other words, a Mike Maples would say, Miller likes to invest in products for which there is such an obvious need that the market will pull it to success. After the fifth or sixth query, Miller realized that the demand for SEO was incredibly large, and he started thinking that he would like to cut himself a portion of SEO market’s financial pie. When he met Kelly, he noticed a few things that he really liked:

1)    Kelly has a “Bias Towards Action”.  He is always working toward the next step, which is always customer oriented. Miller notes that Dell, Apple, and Microsoft were all started by people actually doing things—making products in their dorm rooms (Dell—in fact, Michael Dell, the founder of Dell, is a friend of Miller’s; the two went to college together, and Dell’s dorm room was just across the hall from Miller’s) or garages (Apple), and actually figuring out ways to sell something to people. Amazingly, Kelly does not make balance sheets or income statements or any of “that accounting bullshit.” When I asked Kelly why not, he replied, “As an investor, would you rather me waste my time writing up balance sheets, or actually developing a product that people will buy?” When I asked Miller what he thought about the need for balance sheets, his response was almost exactly the same.

2)    Kelly has an engineering background, and he approaches marketing from a technical, mathematical, scientific, data-oriented perspective as opposed to an intuitive artistic perspective, like “most other people in marketing business” (Miller). Miller and Kelly are interested in what extensive data can tell us about how to proceed, and how we have progressed. Kelly is highly analytic and economical in his approach to solving every kind of problem, and in his approach to designing his product.

3)    Pear Analytics is a scalable business. Because the web analyzer is an automated online tool, the cost of serving 10 customers is not much different from serving 10,000.

4)    Pear Analytics has, as Miller put it, “good traction”, meaning that the business model is subscription based. At the time of investment , Kelly had begun developing a product called SiteJuice, which would be a subscription based service that monitored a website over time. A customer would be able to track the progress of his/her website as he/she made changes and fixed the problems diagnosed by the analyzer tool. The service would be rendered for a monthly fee. All they have to do is make the sale once, and, as long as they deliver an excellent product/service, it keeps on giving. Moreover, since Google’s algorithm keeps changing, Kelly’s customers would need the continued monitoring from SiteJuice to see what aspects of his/her website to focus on at a given time, and to see how effective the changes he/she makes are.

5)    As Miller said, Kelly seemed to “get it.” People often ask, “How much should I spend on SEO?” Kelly can demonstrate the value of his service for each individual customer, based on their website, by actually calculating it. Let’s say you’re a customer asking this question. First you assess your current position in a Google search for a relevant keyword. If you know where you appear in that search, you know approximately how many clicks you will get. Now we find the average search volume for that keyword. We can do this by using the Google AdWord API. (An API is an application programming interface—it is an interface implemented by a software program to enable interaction with other software.) Now we find the average cost per click. This is the price you would pay per click in a PPC paid search campaign for the keyword. Usually it’s around $3/click, but it can range up to $12/click and higher. You can calculate your SEO Value with the following equation: (Position Click Percentage) * (Avg. Search Volume) * (Avg. Cost Per Click) = SEO Value. Your potential SEO Value is calculated by assuming that (Position Click Percentage) = 0.42, to represent the 42% of clicks you would receive by being at the top of the Google search. If you calculate a $100 SEO Value for a certain keyword, and calculate a potential of $1000, you are only reaching 10% of your potential. SiteJuice calculates all of this for you, as well as for your competitors so you can compare and see where you stand.[3]

6)    “Fischer Price Language” made SiteJuice incredibly easy to use. Miller reasoned, like Kelly, that most people setting up websites aren’t too tech-savvy, which is why they need SEO help in the first place. Making the product easy to understand was a key factor in tapping into the market. Moreover, other free website analyzers, like Google Analytics, required you to download and install something on your computer—often a cumbersome process. SiteJuice would require no installation, but would crawl the website each time you asked it to perform analysis.

7)    Pear Analytics is an easy, memorable name. (I will address the name in a later section).

8)    SEO is an incredibly large and growing market. According to a July 2009 study by Forrester Research, search engine marketing was then at $15,393 million—12% of total advertisement spending. It is projected to be $31,588 million—21% of total advertisement spending. (Appendix 2 contains a condensed graph of the Forrester study.) Miller noted that there is a great deal of competition out there in the SEO field, but rather than intimidating him, he views it as an indication of the strength of the market. If lots of people are doing it, they are doing it for a reason.

Miller and Condon’s investment accelerated the development of SiteJuice, which would become, as Kelly calls it, Pear Analytics’ “flagship product.” In the first week of October 2009 Kelly launched a beta version of the product at the Search Marketing Expo in New York. Kelly would later reflect that although the Expo was a success, perhaps he should have waited until the product was further along in development before launching the beta version. (I have my thoughts on this, which I will discuss in a later section). Kelly and his team stayed up until 3:00 am in the lobby of the hotel the night before the event writing the code for the product. Kelly is a self-proclaimed perfectionist, and it frustrated him that the product wasn’t perfect when he presented it to the conference. He did, however, manage to acquire over 200 sign-ups for the beta version with only 14 hours of floor time, which was great. [4]

With over 200 new customers, Kelly began refining the SiteJuice product, as well as developing ways to market it. To the former end, he encouraged his customers to provide constructive feedback, which he took very seriously. He also took a page out of Rackspace’s book and adopted a “Fanatical Support” position. He set up a 24-hour 800-number and a live chat, so that customers could contact him with questions at any time of the day or night. Kelly recalls waking up past 2:00am on a number of occasions to speak with a customer in need.

The emphasis on listening to customer feedback was very important for Kelly, and he learned a tremendous amount from his customers. He found that many people were having difficulty navigating the site itself and couldn’t figure out how to actually use the product, so he paid $72 for three customers to videotape themselves using the product as they narrated their thoughts. This allowed Kelly to see where people were getting stuck, and as a result he is refining SiteJuice to have a 1-2-3 set up that guides new customers through setting up an account. He also implemented “scroll-over explanations” for elements of the website—when your mouse scrolls over a term, a small box pops up that explains what the term means and how it relates to you.

One of the strengths of SiteJuice is that it allows you to monitor multiple sites at once. So a customer can monitor his/her website as well as those of his/her competitors. This way the customer can keep track of what his/her competitors are doing and how they rank in search engines, and thereby figure out how to out-rank them.

Kelly has so far offered his Beta stage users SiteJuice for free, and that seems to have attracted a fairly large crowd of customers. He recently broke the 500 mark last month (December 2009). But, but this month (January 2010), he hopes to start making people pay. How should he price SiteJuice to attract new customers while keeping the old ones?

Pricing

When Kelly sat down to figure out a pricing model for SiteJuice, he turned an eye toward the competition. What price was the competition asking for their services? The two biggest competitors were Hubspot and SEOmoz.

Hubspot is an inbound marketing software system that helps websites optimize their Google status and convert visitors into customers for maximum marketing return on investment (ROI). They are high profile, have won many awards, which they proudly display on the home page of their website, and their blog, webinars, and website Redesign eBook and kit have been featured in the news and on high profile SEO and IT sources around the world. For small businesses Hubspot charges $250/month for their comprehensive services. For medium businesses, $9000/year, and for large businesses, $12000/year.

Hubspot now has 1,750 customers and, in the most recent round of financing, raised $16 million. Hubspot comprehensively leads a customer through the process of SEO and visitor to customer conversion. It has a large team of experts and is dedicated to providing service to customers who may not be as savvy with the internet or technological issues. However, although Pear Analytics does not have the same name recognition as Hubspot, Pear Analytics’ data gathering process, which is crucial for SEO, is far quicker, more sophisticated, and more elegant than Hubspot’s traditional manual approach. Kelly estimates that his web tool cuts about 80% of the time that Hubspot uses to gather data—time better spent fixing problems rather than identifying them. Moreover, Kelly and his team have questions about the sustainability of Hubspot’s business model. The math works out to show that Hubspot spends about $10000 to acquire each new customer, but each customer pays them on average $6000/year. So Hubspot is going to have to cut their cost per acquisition, or start signing up lots and lots of customers very quickly to maintain growth and profitability.[5]

SEOmoz markets itself as a “hub for search marketers worldwide, providing education, tools, resources and paid services to help every SEO be the best they can be.” (SEOmoz homepage).  They also provide consulting services in SEO and internet marketing, for which they charge $80/month.

SEOmoz has similar capabilities offers similar analytical tools to Pear Analytics, but SEOmoz segments the tools into about 20 different, disparate products that work independently; and each tool takes some time to understand and get acquainted with. SiteJuice does everything SEOmoz’ tools do but quicker, and with only one easy to use tool instead of 20.

With these two points of reference, Kelly and his team calculated Pear Analytics’ average cost per customer. Taking into account all of the expenses of daily life—fanatical support charges from the 800 number and 24 hour, charges from using external APIs, the monthly cost of the server—they determined that the average customer costs them $7/month.

Ultimately, with all of this analysis, quantitative and qualitative, the pricing decision was mostly influenced by subjective opinion, and what the team thought people would pay for the product and continued service. They weighed the strengths and weaknesses of their product versus those of Hubspot’s and SEOmoz’, and decided somewhat on a hunch that people are attracted to products under $100. So they decided to set the price for their “Pro deal”, which includes support and analysis for 40 pages of a website, 10 pages of competitor website, 150 keywords, client access, and automated reports 4/month, at $99/month. As well, the team created a Basic package for free, a Personal package for $39/month, and a Business Class package for $299/month. It was cheaper altogether than Hubspot, and they believed it was a superior product with high scalability.

For the beta users, Kelly wants to cut them a deal that is “stupidly good”—one that would be absurd not to take. He plans to offer them the Pro service, which is what they’ve been Beta-testing, for only $99 for the first year. Kelly mentions that Chris Anderson’s book, Premium (which he reverences frequently), Anderson states that when a subscription business goes from free to paying, there is on average a 2% yield rate. So if more than 10 of his 500 customers sign up, then it will be a success. Kelly is shoot for at least 300.

Marketing and Other Products

When Kelly moves out of the Beta stage, he wants to optimize his marketing strategy. Luckily, Kelly knows a little something about marketing. He is the co-founder the Permission Network, organization promoting the philosophy of permission-based marketing as opposed to interruption-based advertising. The idea is we are currently being saturated with advertisements that interrupt our daily lives to promote a product, and they are losing efficacy. Here’s an example: when was the last time I considered a TV add when I went to buy a tube of toothpaste? Perhaps never. When I need toothpaste, I go to the store and I browse my options, and then I make a decision. Kelly believes that the time to make the sale is when a customer has already decided he/she needs a product like yours, but is browsing the marketplace. He wants to make the sale when you are ready to be sold to. But of course, the first thing to do is get some exposure.

However, in order to free up some time for him to work on this, Kelly needed to bring someone else on who could manage the nitty-gritty details, most importantly tweaking the algorithm to respond to changes in Google’s algorithm and customer feedback, or advancements in his own philosophy. In the summer of 2009 he had brought on Romy Misra, a masters student in Industrial Engineering at Texas A&M University (she is now recently graduated—since mid-December 2009). Misra was studying the Google PageRank algorithm (or what was available about it to study), and the netflicks movie preference algorithm—a complex algorithm that determines what kinds  of movies to recommend to users. Kelly got along well with Misra, and was impressed with her attitude and work ethic. He brought her on full-time as the “senior analyst” and assigned her the important task of managing the secret and ever-changing algorithm, as well as the company’s client analytics and SEO projects.

Once Misra was on board, Kelly could devote his time to developing other things. He and his team have now developed a web analyzer widget that people can customize and put on their webpage. It shows whatever information they choose about their website, based on the web analyzer tool itself. For example, if a website is proud of the score it receives in Pear Analytics’ web analyzer, they can put the widget on their page as something like a badge. The widget invites visitors to that site to type in a website of their choice to be analyzed, and the widget outputs a score. The widget includes a link to the Pear Analytics website, should a visitor’s interest be piqued.

The very existence of the widget creates an inbound link to Pear Analytics, which has the consequence of raising Pear Analytics’ Google PageRank. Building inbound links is hard, and the widget is a clever way to do it. As an added bonus/incentive for people to put the widget on their site, Pear Analytics throws them a “kickback”—or a small percentage of the profits generated from visitors who come to Pear Analytics’ site through the widget on their webpage.

The widget has other applications as well, and is a potential product on it’s own. Around the time Miller and Condon became interested in Pear Analytics, a server company called ServerBeach (another of Miller’s babies) expressed interest in striking a deal with Pear Analytics for a license to use the widget for their website in a special way. ServerBeach wanted Pear Analytics to create a version of the widget that also analyzed a websites speed and performance. A hosting company with the capacity to quickly analyze how well and quickly a website is performing is extremely useful. For example, with the widget installed on ServerBeach’s website, a website hosting with ServerBeach could then analyze his website’s speed. A small report would pop up about a minute later with a score indicating how quickly the website loads from the United States and other parts of the world. If the load time is too slow, then the website would have incentive to upgrade to a more advanced package on the server. ServerBeach and Pear Analytics have worked out a deal where ServerBeach would get exclusive rights to this special widget for $10000/year.

In addition to the widget, the Pear Analytics team is developing a web analyzer application for WordPress and the iPhone. Despite some of Kelly’s friends and customers saying that they would gladly pay for these, Kelly insists that they be free because he wants it to be incredibly easy to experience Pear Analytics’ products. His philosophy is to get people to “drink the kool-aid”, and then they will buy from you.

As aforementioned, SiteJuice is Kelly’s  flagship product. His philosophy is that he is building an army of free, enticing smaller products around it so that people will become interested in the tool, and perhaps hooked enough to buy SiteJuice. But there are other revenue garnering products in development. FixMySite, which Kelly hopes to launch in the next few months, will expand on the idea of the free web analyzer with more complex and in depth analysis. FixMySite will crawl a website, diagnose the problems, and then offer to fix each problem for a fee, depending on the complexity of the problem.

Kelly also wants to build a more specific competition tracking tool. He wants to build a product that identifies and analyzes the top three Google links for any given keyword search, so that a company that wants to be found with those keywords knows exactly what they at least need to do to get to the top.

Kelly also would like to enhance his social media. As of now 33% of traffic to his website comes from social media—his blog, his posts on other blogs, YouTube and other videos related to Pear Analytics, Facebook, and Twitter posts. (50% comes from cold research, and the rest from direct/bookmark, which means that people already know about Pear Analytics and type the url directly into their web browser). He believes that social media is becoming increasingly important for increasing online visibility. Google’s new update, Google Caffeine (not yet released), is rumored to more heavily weigh social media pertaining or linking to one’s website. Eventually (hopefully soon), Kelly would like to hire someone to be a “community manager”, whose job it would be to oversee and optimize all of the social media. He would also like to hire an “SEO Evangelist”, who, among other things, would go around the world and speak about SEO on behalf of Pear Analytics at all the important SEO events. This would give Pear Analytics more exposure in the SEO community and would help establish them as leading experts in the field. Kelly has a person in mind, who is already a well-respected expert in the SEO community, and is trying to bring him on board.

Kelly also needs to get Pear Analytics up to the top of the Google searches for keywords pertaining to his website. Kelly does, of course, eat his own dog food. While the top ten keywords that drive people to most sites have the name of the company in it, Kelly has successfully made it so that the top ten keywords that drive people to Pear Analytics do not. (Of course, if you type in Pear Analytics, it is the first thing that comes up. I mean to convey that the top ten most commonly used search terms that bring traffic to Pear Analytics’ website do not have the name of the company in them.) I will speak about this more in depth in the recommendation section.

By the way, Why the Pear?

As an aside, here’s the answer to the company’s most frequently asked question. Kelly likes to say that the pear “represents the fresh, crisp, organic nature of what we do. There’s no ‘auto-pilot’ here, and everything we do requires interpretation and analysis to develop real insights.”[6] Even in the later developments of automating the data gathering process, Kelly remains committed to doing real organic analysis for each individual customer.

The Future

After speaking to Kelly for a long while, I ask him if he is hungry and he replies “Yes.” He escorts me out of his office as we walk past the neon exercise balls and Wii Tennis I admire the laid-back nature of the office environment—something I forgot when I was trying to digest the complicated information Kelly was sharing with me in the back conference room. We go to lunch at Jimmy John’s and talk about life and business and the overdressed woman who drove up to the poor-boy sandwich place in a Rolls Royce. I would later go back to his office and play RockBand on the resident Xbox 360 with him and talk more about Pear Analytics’ modus operandi, but for the moment I wondered what lay ahead for the company.

Right now Kelly is working on many things at once. He is trying to get the widgets up, while also trying to prepare FixMySite for launch in the near future. He hopes to apply for a patent in the near future. He has not done it yet because he believes that it is a better application of his time and limited money to develop the products for the customers rather than deal with expensive patent lawyers. Even though he holds a patent for something already, he was not involved in the process of acquiring it and so is not familiar enough to go about doing it himself. Not to mention that a patent for an algorithm like Kelly’s will no doubt be very technical and will require the expertise of an experienced lawyer.

Kelly’s goal for the far future is to become a “globally recognized leader in organic search tools and services.” Global, because Kelly dreams big, and organic, because Pear Analytics deals only in optimizing performance using organic search and not paid search. Pear Analytics is a young company and has far to go before reaching that goal, but I believe that it is within reach.

My Humble Recommendations for and Reflections on Pear Analytics

I feel uncomfortable offering these opinions, especially because I admire Ryan Kelly and Morris Miller, both of whom I interviewed, for their intelligence and ingenuity. But I offer them humbly, and with respect.

1) I believe that the biggest problem facing Pear Analytics at present is that it is not at the top of the Google searches I conduct for various terms relating to the company. In Appendix 3 I show the results of four simple searches I conducted. For the search term “free seo anaylsis”, Pear Analytics is fifth in the results. Pear Analytics has the advantage that the top three websites have cluttered interfaces ridden with complex language that might scare people off. But the fourth is their competitor Hubspot, whose interface is clean and clear. SEOmoz is in eighth place. If Pear Analytics wants to be the global authority on SEO, it needs to rank number one in searches for SEO related terms. This is no easy task, because ranking higher on Google requires meticulous building of inbound links. But it should be a high priority for Pear Analytics.

2) I think that the Pear Analytics Website could be more easily navigable, and have more self-promoting information upfront about how their products compare to competitors products. The homepage currently reads, “Having trouble finding your website in the search engines? Rank higher and try SiteJuice™ for free, the world’s best SEO management tool.” Instead of saying that it’s the world’s best SEO tool, why not have a link that says, “What is SiteJuice™? Find out why it’s the world’s best SEO tool.” What I’m getting at is that Fischer Price Language is wonderful, but Pear Analytics should be careful not to exclude good content that will show superiority to their competitors because the reason is complicated. In fact, I don’t believe the reason is complicated. The reason is simply that SiteJuice™ is the only webtool that delivers unmatched insight into a websites SEO problems quicker than anyone else. I think they should say that on the website so that customer’s don’t second guess the statement “world’s best SEO tool”. Maybe they should even include a section of the website that talks about what other companies do and outlines why Pear Analytics is better.

Consider also putting a mission statement on the homepage. The one Kelly told me would probably do the trick: “Our goal is to become the globally recognized leader in organic search tools and services.” Perhaps the team could distill this into more Fischer Price Language, but it’s nice to know what a company is all about from the moment you see their homepage. Hubspot’s and SEOmoz’ homepages both contain brief descriptions of their company’s mission statement or of what they do. Pear Analytics should do the same. But of course, be careful not to clutter the site. The Pear Analytics homepage has a great elegance and simplicity, so they shouldn’t mess that up.

3) The website is also a bit confusing in that it does not clearly distinguish between the products it offers. When I’m on the website I don’t know when I’m using SiteJuice™ or when I’m simply using the free web analyzer. Perhaps a simple breakdown of the products offered could be placed in a sub-page labeled “products offered.”

4) I think Kelly and his team should blog more. Specifically Kelly. Kelly’s blog posts are always extremely insightful and interesting, and the blog is a form of social media that brings loads of traffic to the site. The posts also show Kelly’s personality, and they show off his highly analytic, scientific mind. The more frequent the posts, the more loyal customers will check it, and the more random traffic the blog will draw. Perhaps there is an ideal time in the week to blog? Tim Ferris believes the best times to blog are at 7:00am and 6:00pm on Tuesday, Thursday, and Saturday. Perhaps Kelly commits to one of those times and writes a weekly or bi-weekly blog post? Of course, he would have to not sacrifice the quality of the posts for quantity, but as long as he could find something interesting SEO related to talk about, he will probably talk about it intelligently. Perhaps managing the blog—who blogs when and how often—would be the responsibility of the new social media superviser.

5) As an aside, I think that when Kelly brought Luther on, he should have offered him equity. It was a risky move to share the precious algorithm with a computer programmer and ask him to convert his equations into code without offering him equity in the company. Kelly is soon making Luther an equity partner, but even so, if it’s important that the secret is kept secret, there should be strong incentives in place.

6) In the vein of protection of intellectual property, I believe Kelly needs to get a patent sooner rather than later. Secrecy only works for so long, but what if someone develops a similar algorithm and blows past Pear Analytics? Of course, making the algorithm is hard and requires considerable intelligence, SEO expertise and experience, and really good internet intuition, the possibility exists. It would be even worse if a competitor like Hubspot or SEOmoz developed a quick and comprehensive product like SiteJuice™ and knocked out Pear Analytic’s competitive advantage. I realize that it is expensive and time consuming to do it, but better to start now than later.

7) Kelly and his team will be attending a conference in Austin, TX in a few months where they plan to launch FixMySite as well as achieve a number of other development goals. This is an obvious statement, but I believe they should try to dominate this conference as thoroughly as they did the New York one last year. Nothing like an explosive performance to garner signups and to jumpstart the next phase of business.

Life Lessons

These are presented in no particular order.

1) Do not be a perfectionist. This is a big one for me, because I am a notorious perfectionist. Kelly explained that rapid deployment of good code is better than slow deployment of perfect code. In the first place, you can never know what’s perfect until you actually let a customer try the product out. You might think it’s perfect, but then the customer hates it or can’t figure it out, so you have to change it anyway. A good web-based business will put out new code at least every week. This shows productivity and progress, but it also allows for constant and evolving customer feedback. As an example, look at the Search Marketing Expo in New York. Kelly’s product was nowhere near perfect—it wasn’t even presentable until 3:00am the night before the event. But his presence and performance at the event garnered him 200 beta users, which was invaluable in the development process.

2) A problem Kelly identified that he made was that he set his web developers to work on the same code-base at the same time. This caused confusion among everyone because people would unknowingly write over things that people had written, and this caused some stress. The lesson is to set protocols for cooperative behavior. When people work on the same thing, it’s great when they all want to contribute independently, but there need to be rules in place to make for smooth and successful cooperation.

3) Entrepreneurship is often as much about luck as it is about hard work. That’s just the way it works. It was complete serendipity that Kelly met BrandStack who would later introduce him to the two investors in his company. Had that not happened, Kelly said he probably would still be bootstrapping his business with minimal development progress.

4) While many decisions in an entrepreneurial enterprise are made from the heart, data still reigns king. One of the main reasons Miller and Condon liked Kelly in the first place was because Kelly is data-oriented and scientific minded. He takes action based on analytical thinking and good data. That’s how he designed his algorithm, and that’s how he runs his business. And that’s why the entire field of SEO is booming right now in the first place. People are realizing that good data can help you make better business decisions. When Tim Ferris came to speak to us he said the same thing, but my conversations with Ryan really hammered this point home. Data helps people make more informed decisions about their companies. For example, at the beginning of the recession many companies started to cut marketing costs—that’s just what is traditionally done during bad economic times. But if you had data that indicated that certain marketing efforts were generating 150% ROIs, you wouldn’t cut those marketing expenses and you’d find something else to cut back on.

5) As a corollary to 4, it seems to me that every single decision you make has the capacity to affect your business negatively or positively. That’s why running a company is so hard. The decision to use exercise balls instead of chairs affects the culture of your office environment. Is that good or bad? (I think it’s good.) The decision to fire someone, or to hire someone. The decision to blog now or later, about this subject or that one, will affect traffic to your website and how potential customers perceive you. Every decision matters. That’s a scary prospect, but also empowering because everything you do is a calculated move that can help further you along. If you’re savvy enough, and maybe lucky, you can do it.

6) As another corollary, I realized that becoming an expert in something is really not that hard. When I spoke to Miller he told me that if I really put my mind to it, I could probably become an expert in SEO in about 4 weeks. All I would have to do is know where to go and what to read. But what separates experts from entrepreneurs is the wisdom to know when your knowledge is biased and you need to gather some data to confirm or repudiate your expert hunch. (Editor’s note: while I understand what Morris was saying here, it’s not likely that anyone would become an “expert” in a mere 4 weeks.  So much of SEO is related to highly technical areas of a website that evades even experienced web developers. Also, it will take months, if not years, of trial and error to see what works best for various types of sites.  I believe an “expert” will have lots of practical experience applying various SEO techniques.)

7) Culture is key. Kelly explained that he loves the laidback atmosphere of his office. It makes him and his employees like to come to work every day. They sit on bouncy balls instead of chairs. It was a financial decision to cut spending, but it’s also hilarious. When they’re stuck on a project, they don’t have to sit around and pretend to look busy. They can take a breather and play video games. Google requires that for 10% of the workday, it’s employees work on something non-work related. It’s nice to have time to breathe. There are no set office hours. Employees don’t have to be in at any given time or out at any given time. They come in when they want. Kelly said that he doesn’t care what his employees do as long as the work gets done on time. There are no cubicles or divisions in the office—everyone is equal. The emphasis is on collaborative performance and productivity. Instead of meeting, Kelly has team “huddles”, where they all stand instead of sit. In the huddle, the go through three things: News—personal, general; Numbers (more customers, new data, etc.); and quick updates on projects. They discuss sticking points that people are having and brainstorm solutions, and they discuss things that need to be done, and set timely goals for themselves. All of these components work together to create the culture of the business, which is founded on the principle that when people are happy and relaxed they can be mentally sharp and produce good stuff. If/when I start a company, I hope to adopt some of the same business practices.

8) As a corollary to 6, culture is key, but it only works if everyone buys into it. You have to drink the kool-aid and vibe with the culture of the company. Kelly had to fire someone—the only person he fired—because she didn’t fit with the culture. The details are unclear, but three months after hiring her Kelly realized that she was negative and “needed to be babysat”. In other words, she didn’t take initiative on her own. Positivity and taking initiative are two pillars of the culture of Pear Analytics, so she had to go.

Relatedly, Kelly believes that “resumes are pieces of crap.” In a resume a person can sell him/herself to you falsely. He doesn’t like to hire resumes. He likes to higher personalities with special skills. He looks for good attitude and work ethic, and if you have those as well as the skills he needs, then you fit in. I strongly agree with Kelly’s philosophy regarding resumes, because I believe that a person cannot accurately convey him/herself based on a list of his/her accomplishments. Personality and savvy is far more important to me that GPA. That’s why I stopped checking my grades two years ago—because I don’t want to value myself based on accomplishments. It was nice to hear those sentiments echoed by an entrepreneur I admire.

9) It seems to me that Kelly believes that most things on paper are “pieces of crap.” Like balance sheets and income statements. I was shocked when Kelly told me he just doesn’t do them. But it opened my eyes to the notion that an entrepreneur is nothing more than a salesman. He is selling a product or a service, and nothing matters except the completion of the sale. If he can’t close a sale, then he fails. So it makes sense to me that Kelly would want to spend all of his time developing a product that will sell. That is the essence of entrepreneurship.

10) When Miller told me that he liked that Pear Analytics had lots of competition because it confirmed for him the size of the market, I asked him, “What about first movers?” Miller responded by saying that it’s a very large risk for an investor to invest in a first mover because the market is not defined. He’d rather be a “fast follower” in a market that he knows has enormous potential. That’s fascinating to me. He likes to find good markets, and then find or start companies that will disrupt them by the sheer force of their outstandingly better product or service. This is different from what Mike Maples said: “Be different, not better.” Perhaps it’s okay to be better, as long as you are so much better that the market will fold to your hand. It is, of course, hard to determine in the early stages if a company has that market-bending potential, but that is the job of an investor to intuit.

11) Another lesson I learned is that there comes a time when being proactive about getting investors is better than bootstrapping a product. I am skeptical when Kelly says that if Miller and Condon hadn’t come to him, he would still be back where he was eight months ago. If Miller and Condon hadn’t discovered him, he probably would have realized that his business would stagnate and fail if he didn’t get some money fast, which is when he would have decided to put together an amazing pitch and go find some investors. Being proactive about getting investors when the situation is dire is better than letting your company stagnate.

12) Be biased towards action. A strong entrepreneurial philosophy is to get to selling as quickly as possible. That’s the goal. That’s how Dell did it. That’s how Apple did it. And that’s how Microsoft did it. All three entrepreneurs were biased toward action, and that’s how they started some of the most profitable and influential businesses in history.

I pledge that this paper represents my own work in accordance with University Regulations.

~Josh Lavine


[1]Pear Analytics Website, Ryan Kelly’s Bio <http://www.pearanalytics.com/about>

[2] <http://www.google.com/corporate/tech.html>

[3] Pear Analytics webpage, What is SEO Worth? < http://www.pearanalytics.com/what-is-seo-worth>

[4] Kelly wrote about the experience in his blog on the Pear Analytics website, and in the post he discusses his strategy for the event. His goal was to “dominate the event,” and he was successful. He identifies five factors that contributed to his success: 1) Make it uber-simple to sign up for your product, 2) Simplify your give-away prize (sell your product, not the silly prize), 3) Refine your selling points and messaging based on the audience, 4) You don’t need a $4,000 booth set-up, 5) Huddle with your team to refine strategy on the go. (Pear Analytics Blog, page 1. http://www.pearanalytics.com/blog/)

[5] Kelly, Ryan. “Hubspot: High Stakes or High Profit?” Pear Analytics Blog. October 29, 2009. <http://www.pearanalytics.com/blog/>

[6] Pear Analytics Blog, page 2. < http://www.pearanalytics.com/blog/page/2/>

Ryan Kelly

Ryan Kelly

Ryan is the founder and CEO of Pear Analytics and has helped hundreds of customers with their Internet marketing since 2003. He has spoken on various topics of SEO, Analytics and other marketing at conferences in New York, Chicago and Vancouver. Clients he has consulted include Sears, KMart, CareerBuilder and PEER 1 Hosting. Ryan currently teaches two Internet marketing classes at Trinity University in San Antonio.

3 Comments

Leave a Reply