Monthly Archives: January 2009

January 30, 2009

What Makes A Great Startup?

That’s the zillion dollar question.  And no one knows the answer definitively.  Even the most successful VCs have major duds in their portfolios.  But every startup that becomes a large profitable company has the following two elements in common. 

1) Product/service people really want or need

A “product/service people want” is the starting point for any successful startup and part of the reason that I love working with Y Combinator startups.  They drill the mantra “make something people want” into hackers’ heads who are actually capable of executing the vision. 

MBAs often spend way too much time obsessing over the business model before they’ve figured out how to create a useful product. A great business model can never make up for a product that doesn’t meet a want or need. 

I don’t really consider myself an expert on creating useful products.  In fact, I’m not sure anyone is an expert.   Steve Jobs may be considered the world’s best product visionary, but NeXT Computer was hardly a smash hit.  And the executive behind Microsoft’s lucrative Xbox business has added much less value with the Zune. 

I was lucky in my first two startups to work with great products – the original founder’s vision really resonated with users.  I helped both companies reach their potential, but I didn’t create that potential.  Luck of stumbling into great products can’t last forever, so I now obsess over finding better ways to figure out if a product has potential before committing to take it to market.  Every launch program starts with a discovery phase where we dig into how well the product is resonating with users, who really needs it, and why it’s resonating.  Then we decide a timeline for going to market.

The only way to know if a product will resonate is to get actual users on it – and the sooner the better.   If the product isn’t striking a nerve, it’s better to delay an aggressive go to market push.  Many startups succeed with a refined vision rather than their original product.  See this list for examples.  

Sean O’Malley’s blog and Eric Ries’ blog are both great resources for helping you hone your product.  But remember, the only way to know if you’ve succeeded is to trickle some users onto it.  Sean O’Malley’s slideshare presentation below is also very helpful.



2) Business model that works

Ultimately startups get VC funding based on their potenital to create a thriving business.  This requires combining a needed product with a business model that pays the costs of building a lucrative business.  There is as much art in creating a strong business model as there is in creating the perfect product.  It is a thing of beauty when all the pieces fit together in a perfectly tuned economic engine.  Each ingredient is relatively simple, but making them work together at scale is extremely difficult.  

These are the key variables to consider when developing a business model that supports profitable, scalable user acquisition channels:

  • Lifetime value of a user
  • Cost of acquiring a user
  • Marginal costs (besides acquisition cost)

The lifetime value of a user must exceed the cost of acquiring the user and any marginal material/service costs (costs that increase incrementally with each customer).   This is generally pretty easy to achieve if you have low marginal costs.  Most traditional software has zero marginal cost, which is why freeware is possible (it may not be profitable, but it is sustainable).  If you’re lucky, the lifetime value of each user is significantly higher than the marginal cost.  In this case you have a lot left over to spend on profitable customer acquisition.  On the other hand, if you have marginal costs that exceed the lifetime value, then this is a non-starter, no matter how useful the product is. 

If your product is useful and the basic business economics work, then the next part of the business model puzzle is figuring out “customer acquisition channels.”  VC funded businesses must have very scalable customer acquisition opportunities.  No VC is interested in funding a business that maxes out at $1 million/year in revenue – even if it has 90% profit margins. 

Once you have a basic engine that works, keep tuning all pieces to make it work better (improve conversion rates, bring marginal costs down, find ways to increase LTV…).  This will open additional profitable customer acquisition channels.  And obsessively tuning all these areas has been a major factor in my ability to attract 10’s of millions of users for startups that ultimately filed for NASDAQ IPOs. 

The Ultimate Startup

The ultimate startup would be one where the product meets a critical need for a huge addressable market, users have a very high average lifetime value, there are no marginal costs  and there are very scalable user acquisition channels that are completely free  (ie viral).  Unfortunately I don’t know any businesses like this.  Facebook comes close, which helps explain their valuation of $15 billion (who knows what it is now??)…  The only piece they are missing is a high lifetime value per user. 

The science behind viral marketing has rapidly evolved in recent years, so I’m axiously waiting for this ultimate startup to launch. Hope I can get some of the early equity in it.

January 21, 2009

Recruiting Startup Marketers from Wall Street

Rather than wasting their time on Wall Street, Mathematicians should be guiding online marketing for startups. 

For years Wall Street has used brilliant mathematicians to create investment models that they hoped would reduce risk and generate billions of dollars in investment returns. They increasingly leveraged their investments falsely believing that they had eliminated most of the risk – which of course added more risk.  Unfortunately most Wall Street investments are based on speculation which makes it is nearly impossible to remove risk regardless of the sophistication of the model.  Before I stopped watching the news CNBC was blaming these mathematicians for creating the complicated investment instruments that led to the recent collapse – claiming that even the CEOs didn’t understand them. And it’s not the first time that too much trust has been put into the abilities of these whiz kids. The financial crisis of 1998 has also been blamed on overconfidence in mathematicians ability to predict speculative markets.

I have zero confidence in really smart people being able to predict speculative markets. I’ve never trusted mutual fund managers with my cash – instead always putting most non-angel investments into S&P 500 index funds.

However, this is a place where mathematicians can create vast wealth – and that’s in startups.  The returns in online marketing are a lot more predictable than investment banking.  By knowing the lifetime value of your users, you know exactly how much you can pay to acquire new users with an acceptable profit margin.  As long as you don’t saturate a source, it generally delivers the same ROI with each campaign.  The beauty is that a very small investment can give you excellent guidance for the returns of a much larger investment.  Even with 7 figure monthly budgets, I’ve always insisted my teams test every new media with $5o0 buys.  I’ve used this approach to discover ways to spend millions with a very fast return on investment.

At my last long term VP marketing role, my first hire was a trained actuary (the guys that calculate risk for insurance companies).  And the marketers at two startups I’m working with now are both sharp mathematicians – one recently graduated from MIT with a math major.

I first witnessed the power of marketing number crunchers when I was at Uproar.  In 2000 we acquired a startup called iWin.  In a very short time they had created the second most popular casual game website in the world on cashflow positive results.  Their secret weapon?  Several math whizzes in their early 20s who had spent a year in investment banking before running the iWin marketing and product teams.  They were so effective that they took over the marketing and product leadership at Uproar (I had already moved on to President of Uproar Europe).

The math behind viral marketing is even more intriguing.  Read Andrew Chen’s Blog  for the inside scoop on how it works.  Viral Marketing has created some of the fastest growing companies in history and most have never spent a dime on marketing.  And who is dominating the field of viral marketing?  You guessed it – mathematicians. 

Unlike investment banking where leverage increases both risk and reward, in online marketing leverage only increases the reward.  The 12in6 Methodology is all about focusing on high leverage projects that improve the ROI of every future marketing initiative.

Looking to hire someone to lead your marketing?  Hire one of the recently unemployed Wall Street analysts (and show them this post to get them excited about the potential of their new job).

January 20, 2009

The Startup Marketing Launch Process is Broken

**See updates at bottom posted on Jan 20, 2009**

Originally published March 2, 2008

The majority of VC funded startups fail and a large part of the blame should fall on marketing.  Specifically, executing a flawed marketing process during the startup’s critical customer traction stage.

Through running marketing at two startups for the full cycle from launch to IPO filing, I’ve discovered that success at various stages requires very different marketing skills.  It also became clear that early stage marketing execution was the most critical to long-term success.  Yet it is nearly impossible to get good at this critical marketing stage.

Why?  Because effective marketers don’t get enough repetition in the early stage to master it.  Any skills they do develop become rusty.  Stock option vesting periods lock them in well beyond the traction stage (typically four years).

I actually stayed five years in each of my last two startups.  In that final year I had very little time for hands on marketing; I was too busy with such things as managing a team of marketers, recruiting more marketers, meeting with the sales team and other executives, preparing for board meetings, traveling to conferences and trade shows, etc, etc…

I know that my skills are best suited to the earliest stage of marketing, but I wasn’t about to walk away from extremely valuable options.  Even after the options vest it’s still hard to walk away.  Beyond paying hundreds of thousands of dollars to exercise options, you also have to pay income tax on the appreciated value of those options.  If the company isn’t public, you can’t even sell the options to get the money to pay the tax…  Anyway, the point is that despite knowing I’m best at marketing during the early traction stage, I was compelled every year to let those skills get rustier as my options appreciated and vested.

My solution to the problem may seem a bit radical at first, but considering the billions lost in failed VC investments it deserves careful consideration.  Here it is: Startups should plan from the beginning to have different marketing leaders at different stages of the company.  One marketing leader to gain traction and kick start growth, one to manage growth until an IPO and one for post IPO leadership.  Considering the average tenure of a VP Marketing is less than 2 years anyway, this really isn’t that radical.  It’s just planning the transitions rather than making a bunch of disruptive firing/demoting/hiring decisions.

You might be thinking that a consultant approach would work here, but I believe to be effective the marketing leader needs to be totally immersed in the role.   Another common approach is just to force the early stage marketer out when they become less effective (the disruptive approach mentioned above).  If they have played a key role in the company’s success, I don’t believe this is a very ethical approach – even though it’s probably the best thing for the company.

So rather than forcing out the effective early stage marketer, have an agreement from the start that it is a short-term role.   I recommend calling it an interim VP Marketing role and planning for full time 3 to 6 months followed by another 6 to 12 months of advising (working with the longer term VP marketing).  This ensures full knowledge transfer and gives the company access to two sharp marketing thinkers during the very important second stage of the company’s growth.  Options will still be an important motivator for the Interim VP Marketing, but they should have a much shorter vesting period.  The total options allocation to marketers will be higher, but this approach should result in faster market traction, meaning less burn and less need for future dilutive rounds of funding.

It’s probably already clear that I am now specializing in this traction stage.  Xobni is my first assignment.  Of course everybody warns that it will be tempting to want to stay on (especially since Xobni is really picking up steam), but I am very committed to developing this approach over the next few years.

Another advantage of this approach is that it will hone my ability to identify great startup opportunities.   Even the best marketing approach can’t save a crappy idea.  The challenges and opportunities of each former assignment will be fresh in my mind when I look for the next startup to join.   I’ll try to avoid startups with key challenges that I could not previously overcome and try to join startups that have the types of assets that proved important in an earlier assignment.

This knowledge is also very valuable to VCs and I already have several that have asked me to help them assess new investment opportunities.   I’m expecting this will be my pipeline for finding new startup opportunities. Given the alignment of my interest with VCs in picking the right opportunities, they are willing to pay me to conduct a marketing viability assessments to dig into target customer’s need for the solution, real addressable market size and segments and any existing current demand for the category.  If everything looks good after this assessment, the VC can make a less risky investment and I can make a less risky decision to try to take on the interim VP marketing role (if a marketing leader is not already in place).

Update Jan 20, 2009: I temporarily removed this post several months ago with the intention of making a few edits and quickly reposting it.  Unfortunately it slipped through the cracks despite being one of my more popular posts.  My thinking has a evolved quite a bit since I wrote this post 9 months ago.  During that time I have nearly doubled my experience taking startups to market (despite being in startups for 10 years).  As much as the idea of interim VP Marketing roles sounded good at the time, it really limits my ability to help several startups and requires more energy than I could possibly muster (this is a very intense period in startups).  Instead I have shifted my focus to work alongside a long-term marketer and guide them through executing the key phases of going to market.  This approach has worked very well at both Dropbox and Eventbrite.

We still have a long way to go before the launch problem is fixed at VC backed startups, but there has been a lot of progress in the last year.

January 19, 2009

Update to 12in6 Methodology Presentation

Here are the latest updates to my presentation on Slideshare giving an overview of my go to market approach. I simplified the overall presentation and contrasted the 12in6 Methodology to the typical approach taken by startups.

Startup Launch – Key Steps to Success

View more presentations from Sean Ellis.

For those who are new to the Startup-Marketing.com blog, this is the approach that I’ve used to launch several successful startups including two that have gone on to file for NASDAQ IPOs (Uproar in 2000 and LogMeIn in 2008 – pending).  Recent startups using the methodology have included Dropbox (runner up for best startup in 2008 at the Crunchies), Xobni and Eventbrite.

Looking forward to any feedback.

January 12, 2009

6-Month News Vacation

I’m a news junky and have been since college.  Recently I’m finding the damage of paying attention to the news far outweighs the benefits. 

For the past two weeks I made a concerted effort not to read or watch the news.  By this past Friday night I had reached my most optimistic outlook in years.  The companies I helped take to market in 2008 are performing beyond my wildest expectations.  Earlier in the week Xobni raised a $7 million Series B round and that evening Dropbox had been awarded runner up for best startup in 2008. 

My H1 2009 workload is quickly filling up with fantastic group of well-funded startups.  And most important – I’m really having fun helping startups figure out how to drive massive customer adoption.  Through it all, I’ve managed to spend more time with my kids than at any other time in their lives. 

What could possibly screw up this optimistic mood?  The news.  I woke up Saturday morning and decided to check in while I drank my coffee.  Big mistake.  After a few minutes of gloomy economic reporting, murders, and war I felt the pessimism creeping in.   Then I picked up the remote and turned it off. 

I decided I’d give it a break for 6 months.  I’ll bet that I won’t even know there is a recession if I don’t watch the news.  On July 11th I’ll check back in and see if there is any sign of the recovery that economists are predicting in H2 2009. 

January 3, 2009

How to Determine the Optimal Price for Your Web Service

For the startups I help take to market, one of our most important projects is determining their optimal price.  Unlike companies in established categories with high unit costs, optimal pricing for a software startup mostly relates to maximizing revenue.  An optimal price allows the startup to grow at the fastest possible rate by maximizing profitable investments in customer acquisition programs and/or offering a free version to drive broad user adoption. Considering most software startups simply guess a price, determining your optimal price can become an enormous competitive advantage.  

The optimal pricing project is part of the overall “optimization phase” I describe in my metrics driven go to market approach presentation. 

There are three key factors to consider when determining your optimal pricing:

  1. Price sensitivity– You want to find the price that generates the highest yield per 1000 trials (or visitors, DLs, etc.).  You can find this number by determining how many units you would sell at each price.  For example, if you have a 10% conversion rate at both $8/unit and $10/unit, then $10 is obviously the better price for you.  But let’s say at $20/unit demand drops to 8%.  Despite lower demand, yield is higher at $20 so it would be a better price than $10 ($1600 per 1000 users at $20/unit compared to only $1000 per 1000 users at $10/unit).  I estimate max yield pricing first through surveys and then through experimentation at several price points.  Around launch your volume will be too low for a meaningful sample size, so be sure to launch with “introductory pricing” which should be at the low end of your expectations.  Adjust the price when volume allows you to hone in on the optimal pricing.
  2. Marginal cost– For web services it’s important to understand your cost per unit to avoid pricing at a loss.  This marginal cost is essentially a floor on your pricing.  If you have bandwidth and storage costs that are $5/user/year, then your business would not be sustainable if you priced your service at $4/user/year.  For most downloadable software, there is no marginal cost per user (beyond marketing costs).
  3. Growth strategy– I generally prefer one of the following pricing strategies for innovative products.  One is a Market Builder pricing strategy where the majority of your users are coming through your demand generation initiatives.  Demand generation is expensive (unless driven through viral tactics) and therefore requires premium pricing to create a high allowable user acquisition cost.   An example of a company that took a Market Builder approach to grow the personal remote PC access category is GoToMyPC, which combined premium pricing with aggressive radio demand generation.  An alternative strategy is a Market Drafter pricing strategy.  Freemium pricing is ideal for a market drafter.  Essentially as the Market Builder creates awareness for the category, the Market Drafter swoops in and offers a much better deal (SEM is a good place to focus for a Market Drafter).  This strategy only works when a Market Builder is aggressively investing to grow the category.  I prefer the Market Drafter position when possible (see this post for more details on why).  In the long term, the Market Builder must focus on differentiation to justify its higher prices (or reduce prices)

Once the optimal price has been established, there are many tactics that can used to boost response rates.  These include:

  • Setting the price a bit higher than the optimal level and then frequently discounting it.
  • Using a decoy super premium version to make the version with the “real price” seem cheaper.

My favorite pricing model for driving demand is Freemium, combined with carefully researched max yield pricing on the premium version of the product – then applying the response boosting tactics listed above.   An insightful read on Freemium pricing is Josh Kopelman’s post “The Penny Gap.”  It is an exploration of the “power of free” in driving customer adoption and suggests that elasticity of demand is not linear.  At the price of zero, demand soars. 

Dan Ariely also makes this point in his book Predictably Irrational.   He concludes “Zero is not just another discount.  Zero is a different place.  The difference between two cents and one cent is small.  But the difference between one cent and zero is huge.”  He supports this point through the following experiment:  He first offered a Lindt Truffle for 15 cents and a Hershey Kiss for one cent.  Participants (who could only select one) purchased the Lindt Truffle 73% of the time and the Hershey Kiss 27% of the time.  When they were both discounted an additional penny (making the Hershey Kiss free), demand for the Hershey Kiss shot up to 69% and demand for the Lindt Truffle dropped to 31%. 

There are several other great pricing psychology nuggets in Predictably Irrational; I highly recommend reading it.  It goes well beyond the three basic pricing factors presented above.  Some useful points include:

  • A higher price not only positions your product as superior, people may actually have a better experience using the product.  He presents a fascinating experiment that shows people got more relief from a $2.50 pain killer than a 10 cent pain killer, even though they were both just vitamin C.  He concludes “the perception of value, in medicine, soft drinks, drugstore cosmetics or cars, can become real value.” 
  • When we encounter a new product, we accept the first price that comes before our eyes as the anchor.  This price has a long-term effect on our willingness to pay for the product from then on.  He uses the example of black pearls.  Initially there was no demand for them, but when they were anchored to the finest gems in the world with premium pricing, demand shot up. 
  • Differentiation gives more flexibility to increase price.  His example here was that Starbucks differentiated the coffee shop experience allowing them to more than double the price of a cup of coffee compared to Dunkin Donuts. 

Finally remember that technology prices tend to drop over time.  Keep this in mind when determining allowable acquisition cost based on a user’s lifetime value.  Lifetime value will probably be lower when considering future pricing pressure.  It’s better to be ahead of the curve in driving prices lower, which often requires innovation that allows you to profitably offer the service at a lower cost than competitors (for web based services with marginal costs).