Category Archives: Analytics

January 11, 2013

Are Marketers Now Required to be Engineers Too?

As much as I hate to admit it, online marketers with engineering backgrounds often have a significant advantage over non-engineering marketers.  They are simply much more capable of “getting stuff done.” The rest of us waste a lot of time and creative energy figuring how to get the obvious stuff done.

Not only are engineering marketers more capable of getting the essentials done, they can use their reserve of time and creative energy to be scrappy about building marketing experiments.  And of course the experiments they build on their own can be much more interesting than us non-engineers.

One way I have worked around my engineering deficiencies has been to hire the skills onto the marketing team.  For example, in my last long-term VP Marketing role I hired a front-end designer/engineer to design and code landing pages and a dedicated DBA to build reports and run ad hoc queries.

Tools are Leveling the Playing Field

Fortunately the playing field has begun to level in recent years so that non-engineering marketers can be much less dependent on engineering help.  For example, I now use KISSmetrics to build my own reports and run ad hoc queries.  And because it is so easy, I spend a lot more time digging into things like event-to-goal correlations.  This helps me know the events I should be promoting to website visitors in a given lifecycle stage. In the past, I was just seeking visibility into the marketing campaigns and landing pages that were and weren’t working.  KISSmetrics allows me to get much deeper and identify additional levers to improve results.

I’m also no longer dependent on front-end design and development help.  Not only can I build extremely effective landing pages with Unbounce, I can also very easily set up and manage A/B tests. This takes me beyond the stuff I even bothered asking for in the past.  For A/B testing landing pages, I previously had to replicate every ad to point to different landing pages.  It doubled the amount of time it took to set up any new ad campaign.  Now with Unbounce it is all automatic.

Even more empowering, I can actually manage all the content and design updates for our entire website using Optimizely.  This proved to be very useful a few months ago when Qualaroo lacked a front-end developer.  Despite limited HTML skills and no design talent, I was able to make all the necessary changes and still keep the key design templates that were custom created by a talented designer.  Finally, Optimizely made it a no brainer to retain the old version and run it as an A/B test.

What’s Next?

I’m very excited about the range of tools that will emerge over the next few years to further empower marketers.  Eventually I see a time where an individual marketer will have the ability to identify and control all of the most powerful levers for driving growth, even from deep within their website.

An important step that is also taking place is automation of the marketing programs that work.  All of this will continue to free up more time for creative marketers to build marketing experiments that truly push the envelope on results.

June 29, 2010

Deconstructing Startup Growth

Elements of a startup growth curve

After product/market fit, driving sustainable growth is probably the most important/difficult part of creating value in a startup.

For most of the last 15 years of my startup experience, I’ve been the point person responsible for primarily one thing: driving growth.  Even after two IPOs, I didn’t really have a firm grasp of the essential elements of driving growth.  My view has evolved from externally focused metrics-driven marketing, to a more holistic approach built on a solid foundation of product/market fit.

Growth Foundation

Even the greatest marketers can’t sustain growth on a weak foundation.  Eventually, their growth curves crater.

So what is required for a strong foundation?

Must Have Product

The most important element is having a large percentage of users who consider your product a “must have” (over 40% is a good benchmark).  This gives you two key benefits:

  1. The first is that your churn will be relatively low (if it’s a “must have” why would users leave?), so you won’t be wasting resources filling a leaky bucket.
  2. The second is that “must have” products generally maintain strong word of mouth.

Together, these two elements give you a steady upward trajectory of your growth curve until you reach market saturation (hopefully you are in a big market!).

Must Have is Perishable

An important caveat is that your product will stop being a “must have” if a competitor offering a viable substitute enters your space. If they are really a good alternative to your product, then you’ve been downgraded to a “nice to have” and your foundation starts getting shaky.  Therefore, once you become a “must have” it is critical to get to the growth phase of your business as quickly as possible.

Check out my earlier post to determine if your product is a “must have.”

Conversion Optimization

Your ability to accelerate growth will be greatly enhanced if you optimize conversions.  There are many ways to define a “conversion” but for me, it’s a person who reaches the “must have” experience.  If 1000 new visitors come to your website and only 50 experience the “must have” benefit, it’s very difficult to efficiently grow your business.   However, with focused attention on fine-tuning the first user experience, startups often see a 2x – 10x improvement in conversions.

This immediately enhances your growth curve since word-of-mouth referrals begin “sticking.”  It also greatly enhances your ability to find viable, scalable ways to grow your user base (especially when combined with a good monetization approach).

Driving Growth

Most startups entering the growth stage obsess too much on finding a VP marketing capable of building and managing a large marketing organization.  At this stage your more immediate challenge is finding sustainable, scalable growth drivers to augment the organic growth achieved through solid product/market fit and conversion optimization.  If you are compelled to bring in a VP Marketing at this stage, make sure he/she has a track record of developing scalable growth drivers and is willing to make this their core focus until it is figured out.  Otherwise, I recommend instead bringing in a scrappy growth hacker to generate a strong flow of ideas for experiments that will scale if successful.

The faster you run high quality experiments, the more likely you’ll find scalable, effective growth tactics. Determining the success of a customer acquisition idea is dependent on an effective tracking and reporting system, so don’t start testing until your tracking/reporting system has been implemented. Once scalable growth tactics are developed, then a VP Marketing may be important for building and managing the marketing team that will execute these tactics.

One benefit that is emerging from advising multiple startups is that our rate of collective discoveries is accelerating across the non-competitive network of startups. With sharp, creative growth hackers in each startup we are able to brainstorm and test many more tactics.  The best ones are exchanged across the network for everyone’s benefit.

Growth

As the preceding paragraphs hopefully demonstrate, growth is a function of multiple factors.  Focusing on the right factors at any given time offers the best chance of ultimately becoming a high growth startup.  One exception to this rule are startups like eBay, Facebook, and Twitter, where “must have” status could only be achieved after critical mass.  In these startups, they did not have the luxury to focus on one element at a time – instead they had to work on the full growth ecosystem at one time.  But for most startups, you will approach your full growth potential by obsessively focusing on the most important goal for your particular stage.

April 30, 2009

Fast Vs Careful Decision Making in Startups

Reversible or Irreversible Decision?

Fast decision making is often the mark of a great entrepreneur.  But as Steve Blank points out in a recent blog post: decisions have two states: “those that are reversible and those that are irreversible.”  Entrepreneurs should take the time to make careful decisions when they are irreversible (such as accepting money from a VC).  For reversible decisions, he recommends starting “a policy of making reversible decisions … before a meeting ends. In a startup it doesn’t matter if you’re 100% right 100% of the time. What matters is having forward momentum and a tight fact-based feedback loop (i.e. Customer Development) to help you quickly recognize and reverse any incorrect decisions.” 

This is awesome guidance considering the countless hours I saw wasted at my first startup where people debated decisions that had little impact on results.  On my marketing team I quickly ended these debates with “test it.”  When debates extended across departments I abdicated the decision to others but measured the results to make sure they weren’t negatively affected.

Higher Velocity Testing Better than Perfect Certainty

At LogMeIn (my second startup) the goal was to start a testing and analytics culture on the marketing team right from the beginning.  Rather than hiring someone with a traditional marketing background, my first marketing hire had an actuarial background (the people that assess insurance risk).  Next we hired a super fast web designer/developer and an equally fast copywriter.  This team was able to rapidly iterate everything to determine combinations that generated optimal conversions. 

One warning before hiring a math wizard to lead your analytics is that they will often want sample sizes that almost completely eliminate doubt that you are making the right decision.  With the volume of users at most startups, this would limit you to very few tests.  When I suggest the following mental exercise to a mathematician, they usually come around to high velocity testing with lower certainty.   I suggest that they try modeling the results of 25 tests with 80% certainty compared to 5 tests with 95% certainty.  I also explain that we can always go back and test it again when we have higher volume.

Understanding User Motivations

But some decisions are a lot harder to test and require more up front traditional research.  For example, when trying to understand the motivations behind users’ actions (or lack of actions) I generally interview and/or survey them. But as Robert Cialdini points out in his latest book: “We know that people’s ability to understand the factors that affect their behavior is surprisingly poor.”  So in the past, this research often confused rather than enlightened me. 

It wasn’t until I read Four Steps to the Epiphany that I realized you could take a more scientific approach to understanding user motivations.  Steve Blank recommends starting with hypotheses around the key factors that will be important for building your business – such as the real problem you are solving and the people who are most motivated to solve this problem.  By engaging prospective and actual users you can validate and/or refine these hypotheses.  Unlike the previous approach to surveying, we now gain clarity with more user input. 

Still I know there is a lot of room for improvement in my scientific approach to understanding user needs and motivations, so I recently brought on someone to help me take my research and analytics to the next level.  He is Molecular Biologist as well as an entrepreneur who earlier in his career spent 10 years in biotech research.  It should be interesting to see what happens when he applies his rigorous research approach to customer development.  He’ll be joining me for my two projects that start next month.

At early stage web startups we have the distinct advantage over established companies of starting with a blank slate, making it possible to set up much more controlled experiments.  In addition to making better use of tight startup time and money, we also hope to leverage the blank slate to challenge some long held marketing beliefs regarding what really works.

March 23, 2009

Iterating Without Understanding?

It seems there are two camps of “evolved” marketers these days. One group recognizes that it is critical to understand customer needs by engaging them at every opportunity. The other group is completely focused on metrics driven iteration. Until recently, few combined these powerful forces.

I started in the camp of online metrics and scorned the beanbag marketers who didn’t “get” analytics. At Uproar in the mid to late 90s, metrics were our competitive advantage. We tested, measured and optimized everything. We knew we couldn’t afford any waste if we were going to have a chance to beat the heavily funded Silicon Valley gaming startups and the established companies getting into online games (Microsoft, Yahoo, Sony). Ultimately, this obsession with leveraging metrics to track ROI and improve conversion through iteration was key to becoming the worldwide leader in online games and peaking at a billion dollar stock market valuation. Despite their much deeper cash war chests, the beanbag marketers couldn’t compete with our no waste metrics driven approach.

Today the Darwinian economy has killed off most web businesses that don’t leverage metrics, so this is no longer a competitive advantage – it’s a necessity. But many web marketers stop there.

In my next startup I was fortunate enough to have a venture capitalist who helped take our approach to the next level. We attracted his investment with our metrics driven online marketing approach and then he quickly improved it. He constantly grilled me with the question “Who is your customer?” During our weekly meetings he never failed to ask about the last time I spoke to a customer. I got extra brownie points for meeting with customers in person. To be honest I initially focused on engaging customers just to appease this VC. But it didn’t take long until I was able to use this information to improve results. Informed iteration helped us increase purchase transaction rates 10X in just a few months, which made scaling a profitable marketing spend infinitely easier. Later customer engagements uncovered revenue opportunities we never could have found through metrics driven iteration. These revenue opportunities eventually accounted for more than half of the company’s overall revenue volume – making possible the eventual IPO filing.

It wasn’t until I began the Interim VP Marketing role at Xobni that I discovered Steve Blank’s The Four Steps to the Epiphany. This book added a systematic process for uncovering the critical information needed to build a thriving business and keep improving results.  The great news is that Steve Blank recently started blogging at steveblank.com. Perhaps even better news is that Venture Hacks now records Steve Blank’s lectures at UC Berkeley and posts them online.

The same Darwinian forces that made metrics a necessity for online marketers are once again shaking up the web startup world. It has become a major competitive advantage to combine Steve Blank’s customer development approach with informed metrics driven iteration. And it’s only a matter of time until this approach becomes a necessity for survival.

So what’s next? I’m certain that eventually a platform will emerge that ties it all together. This platform will facilitate the process of collecting and analyzing actionable customer information and manage the iterations that deliver optimal results. Up to this point we’ve always had to custom develop these tracking and reporting systems, while using disconnected systems to drive understanding (surveys, Excel…). Off-the-shelf analytics programs have been bloated with data that is useless for improving results.

Rather than holding my breath for someone to deliver this dream platform, I’ve been advising KISSmetrics as they work to create it. I’ve given them total visibility into my approach and turned over reports that have evolved over many years of execution. Of course they have given me equity in the company – but I’d be passionate about this metrics driven customer development platform either way.