MicroConf 2011 – Sean Ellis

By , August 10, 2011 3:00 pm

Sean Ellis has a successful track record launching and running several web based businesses. Sean talks about the various companies he has worked with and the strategies and tactics that have served him well. Sean also talks about the new and different challenges that he faces with his new startup www.catchfree.com

Sean blogs at www.startupmarketing.com

Are funded companies different to bootstrapped companies?

All startups have a different approach to market. Sean has had different experiences with the various companies he has run.


Sean’s first company was Uproar.com, based in Budapest, Hungary. Uproar was an online games company. Sean’s lack of marketing background was an advantage for an internet based company early on (1996).
They had a limited budget, so they spent effort on efficient customer acquisition efforts:

  1. Could not compete against well funded company
  2. A/B tested everything
  3. Tracked return on investment on all money spent and constantly shifted costs to match
  4. Focused on purchasing CPA not CPC, mainly done using viral marketing


Sean’s second copmany was logmein also based in Budapest, Hungary

Created logmein because founder already had experienced the need for this product in a different business model. However, by the time they got to market, they could not compete with well funded, high traction GotoMyPC, so changed to a freemium model to disrupt and differentiate against them. Logmein is now a $billion business listed on NASDAQ.

  1. Differentiate product against competitors.
  2. Freemium model.
  3. Harvest demand, use SEO to draw customers in.
  4. Understand customers, could find extra use cases for your product (new markets!). Survey customers. Reassure any concerns customers have.
  5. Optimize funnel. Done in house, but products like KissMetrics do it for you.
  6. Customer surveys, built into the product and built into the website.
  7. Simplify all “hard” problems.

Sean has also used these strategies at xobni.com, dropbox.com, eventbrite.com, lookout.com, socialcast.com.

Sean thinks you should double down on any area that shows potential. Focus on the people that like the software. See if you can improve that and drive even more conversions. Focus on that, not on the “why don’t these people…”.


For the types of business above Sean has developed a framework to explain the business. The framework consists of these key points.

  • Great product
  • Proiritise / focus
  • Leverage most important aspects first
  • A/B test sales funnel
  • Build in viral loops

The pyramid describes the stages of the business growth.

  1. Product Market Fit – ensure you gratify your customers.
  2. Transition (efficient conversion).
  3. Scale – this should only be attempted once the previous two steps are complete.


You can validate your ideas about the product by asking your customers this question. “How would you feel if you could no longer use this product?” then categorise the answers according to how many people are very disappointed they cannot use the product.

Very disappointed Action
0..25% Keep burn low, engage, iterate. This is an OK place to be if you are not trying to grow business.
26..39% Try repositioning the product, retargeting your customers
40..100% Proceed up pyramid

Race to Scale

To be able to scale the business you need to understand the core value perceived. Focus on value added products.
Position the product, based on value to the customer.

Example: If 7% of customers like it, focus on the things they like and remove the others, change the messaging, test it, maybe you’ll get > 40%. Maybe the issue is just highlighting the right things in the message

It is important to measure the right Metrics, focus on the right thing:

  1. Value of customer
  2. Cost of acquisition
  3. Up selling/Cross selling.

Optimize conversions.

Realize that at some point A/B testing will distract you from improving other aspects of the business.

So A/B is testing is important but do not get obsessed. At some point the effort required to get an extra 1% conversion rate is not worth the return you can get by optimizing some other part of the business.

You need to have viable economics before going for growth.
Optimise each step of the funnel before growth.

Network effect business

Sean’s latest company is not like the other companies he started. Sean wants to try to help businesses working with the freemium model get more exposure.

The framework described earlier does not work for this new company because you can only get growth if you have customers but you can only get customers if you have products to promote and you can only get the products to promote if you have customers. A classic Chicken and Egg problem. Sean describes how he is dealing with this problem.

Problems with Freemium:

  • Skeptical users, why is this free?
  • Awkward economics, how to generate revenue from free?

Catchfree solves this by providing a place for people to vote on their products. Once you have customers you can provide CPA (cost per acquisition) advertising rather than CPC (cost per click) advertising.

The diagram for this business is the reverse of the previous business model.

With this model you start with lots of customers and lots of products and only at the end do you find out if the idea works. Its completely backwards compared to the previous model where you can measure, predict, test, change everything up front.

The implications of this are that:

  • The business needs embedded tracking to handle CPA.
  • You need to validate demand before you do anything.
  • You will get either a lot of funding from investors or none at all.
  • You need to define a minimum viable critical mass of customers for the business to work. What is that value? Is it 100?, 1000? 100,000?, etc. If there are not enough customers other customers won’t have a reason to beleive in the business and there will not be enough revenue to make the business work.
  • You need to have really clear rollout goals.
  • Spend time getting architecture right before launch – due to the number of users required to make the business work this is an important factor. Sean has opted for an agile development environment.

Iterate/optimize on the fly.

Sean recommends targeting the minimum market you can to start. This reduces costs and allows you to optimize the impact for that market. Sean used twitter to target initial market because they match the initial profile for catchfree.

You need to generate revenue early on. Catchfree are not spending much to get customers.

After launch, you can validate customer gratification. Extend targeting and marketing based on customer validation.

Catchfree uses a reverse affiliate model. They give CPA credit to vendors based on customer recommendations. If a customer recommends their product the vendor gets a credit against future CPA costs. This was catchfree don’t have to spend money on credits but vendors do receives something useful (free CPA which they use to increase their exposure).

A lot of marketing is done via sharing (viral), tweet and share on facebook. Sean recommends Andrew Chen’s blog.

After getting search optimized by working on SEO and SEM the last thing to do is demand generation (adverts not on search engines, ie on TV, radio, magazines, etc).

Sean’s final recommendation is when rolling out, always start with the free stuff first then the low cost stuff. Get the maximum impact you can while spending the least you can. After you’ve done free and low cost things can get expensive.

The linked audio file has some questions at the end:

  • 52:30 “Clarify reverse affiliate”.
  • 53:50 “Funnel optimization”
  • 56:15 “Funding vs Unfunded, different revenue patterns”

Tommorow I’ll be posting Rob Walling’s talk.

MicroConf 2011 – Andrew Warner of Mixergy

By , August 9, 2011 1:11 pm

The first speaker at MicroConf was Andrew Warner of Mixergy.

This write up is my take on Andrew’s talk. I’m listing the things Andrew picks up on in pretty much chronological order as he talks. If anything doesn’t quite make sense I encourage you to listen to the MP3 shown below.

Andrew talks with passion about what drives the entrepreneurs he interviews. What makes a business bomb and then the next business with the same people is a roaring success. Andrew gives some examples: Andrew Mason founded The Point which bombed but also founded Groupon which is a success, After AirBnB flew to New York their business takes off – why is that?

Andrew is very curious about what makes a good business. Andrew interviews a diverse group people and on many subjects but a common thread comes from them all.

How to stop making mistakes? Surprisingly, we are more likely to make mistakes if we know not what to do. How can we mitigate this? To illustrate this Andrew’s talk gives examples of various different people and what they did and the results they got or the strategies they use.

Hugh MacLeod – Why does he draw on back of business cards? Convenient, no need to get to work, buy a canvas, not confined to one physical place (where the drawing board/canvas is). Business cards, can draw anywhere, practice any time, any place. Practice makes perfect. Repetition allows improvement by testing and allows you to repeat until you find what works well.

Gary Vaynerchuck produces wine tasting videos, every day. This allows him to refine the process. He decided to leave the mistakes in the videos – shows you are human (and saves time not editing them out).

Andrew sometimes asks awkward questions. Sometimes that goes wrong, but often people answer unexpected questions, questions you think they would reject. This allows more insight. Its good to push the boundaries sometimes. Do not be afraid.

AirBnB – Why did they go to New York?
They were struggling to make the idea work in San Francisco. So they took some advice from Paul Graham of Y Combinator. . Most of AirBnB’s customers are in New York. Paul Graham’s advice was “Go to New York and talk to your customers. Don’t care about scale. Just do the talk.” They recognised key insights from talking to customers. They thought adding photos would be good, but it turns out the photos were not big enough. Until you get the feedback you only have your own frame of reference – people want unexpected things. Then they also met a musician (Barry Manilow’s drummer) who wanted to rent his whole place while on tour, not just a room. So changed their business model based on this feedback.

Small insights – big changes.

Fusion Charts is multi-million dollar company selling charting and graphing software for use on websites. Their first customers wanted tech support for their graphics widgets on website. They charged peanuts to provide tech support. Why? So that they could get in their system and see how they are using their software – after a while of doing this they worked out how to simplify and scale it.


Avinash Kaushik is a noted expert on web analytics. He likes AppSumo because of the testing mantra. He mentions Noah Kagan’s sig line (which is itself an experiment – try it – I tried to find an example of this but it seems I have deleted any that I had, sorry).

Andrew’s business Mixergy sells access to interviews. Andrew tried selling lifetime membership just to see if people were interested. Andrew is constantly testing new ideas.

To understand success you must study success AND failure.

fitfuel.com started off selling nutritional supplements. Then started adding energy bars, and cereals. And so on until he ended up with health food for dogs and viagra (sexual health). He ultimately ended up selling viagra for dogs. All the changes were test driven, but there was no focus. The tests worked but the result was diffuse. He had too much stuff in the warehouse going out of date. That is effectively money going rotten. The result? – business failure.

Techstars believe in power of small experiments. One of the startups they helped Photobucket is a photosharing website. They allowed embedding on websites. They killed ability to embed and customers complained. So they brought it back. Kill stuff to see if people complain. Andrew tried the same technique on mixergy by removing transcripts. Not enough complaints so dropped the transcripts – resulting in less work and a better, tighter experience.

Tim Ferris is author of The 4 day body and 4 hour workweek. The claims seem outlandish, but underlying them is the concept of “Minimum effective dose”. How little do you need to do to get good results? Seems that many entreprenuers are like this.

Ramit Sethi is author of the book “I will teach you to be rich”. Andrew shows the screenshot of a website. It has info graphics, embed and include anchor text, and a bit of SEO with the embed text. Outsize results with small changes. (I’m afraid one is too visual, so you’ll have to take my word for it).

Artificial demand

An alternative way of doing things, rather than allow people to give you their email address is to ask them “request an invitation”. This technique works due to the power of persuading someone to do something. By asking, they are showing interest. If they are interested then maybe they might be interested in your business at some point.

Neil Strauss, author of The Game was continually testing ways to attract girls. He eventually learned what works – “Cat-string theory”. String is interesting if the cat can’t have it. String is not interesting if cat can have it. Neil concluded that girls are the same. Say hello, then back off, see if they are interested. Apparently it works!?

A real estate company was failing but then they changed business model to average price for all apartments regardless of cost. With a twist. The offer was for a limited time only. The business very quickly became a success.


If you want to go from small to big, need big army to help you. Email can help.

Clocky is a mobile alarm clock that runs away when the alarm goes off. It’s gauranteed to get you out of bed! The creator generated a huge email list while the product was in development (2 years, China) and when launched she had no difficulty selling the alarm clock. The company was profitable in 2 months.

How do you build a 16 million strong email list? Use every opportunity to get people to sign up. Make the site so that people can’t do anything without signing up. Some people will leave (they weren’t really interested) and some will sign up.

WooFoo was a blog, with code how to develop better sites. They slowly built an email list. When launched they got slated in the TechCrunch comments, but their fan base stepped in to protect them. Having 3rd parties respond to negative comments is powerful.

Ramit Sethi when asked “Whats the one thing you wish you had done? ” answered “I wish I had signed more people up for my email list, and earlier too.”.

Why did The Point fail and Groupon succeed?

The Point’s business model was I will do “action” if “condition”. They failed because trying to do too many things at once plus they had too much money up front.

Groupon started with a WordPress widget that allowed an offer to be shown. They used an email mailing list. Its boring, old, but it works. They were 5 offers before someone would buy.

Mixergy. Andrew wanted to test a headline. How do I do that? Via email, 50/50 split. He also tried that with Noah Kagan for a Mixergy party plus tickets. Approached all people via email, before public launch. Evenbrite: Minimum is $1 plus $1.06 fee -> $2.06. Try, free section + VIP section ($25). Tested limited seats. By time ready to post about the event, he had sold out of seats.

Its a mistake to grow too big while not recruiting army of fans.

Noah started selling a video in realtime while demoing stuff to Andrew. Andrew didn’t like it, but it worked, lots of people liked it. The data in the video was more important than the quality of the video.


Grasshopper is a virtual phone system for entreprenuers. They have an offer, tell a friend, get a price break. They get a 15% conversion rate. Seemed like a good idea? So they tried another idea, but it failed. Ruined $500,000+. Scrapped after 3 months, rebuild. Then scrapped again. Reputation can prevent you from trying new ideas.

Sizzle It were using video to promote products but they wanted to do more. There was a lack of focus giving a diffuse result. Growing too large too fast. They decided to change their focus to one narrow niche. 1 minute videos about products. Success!

The founder of fitfuel.com (which was mentioned earlier as a failure) tried again by focussing on Christian based nutritional supplements. Focussed, narrow niche. This time he succeeded.

At the end of the talk Andrew takes some questions from the audience. Sean Ellis asks “Who would Andrew invest in?” Answer: “Noah Kagan”. He Loves The Copy.

Tomorrow I’ll be posting Sean Ellis’s talk.

MicroConf 2011

By , August 8, 2011 12:50 pm

Two months ago I went MicroConf in Las Vegas. MicroConf is a conference for the self-funded statup, organised by Rob Walling and Mike Taber.

I always said I’d never visit Las Vegas but I decided pretty quickly that I’d have to change my mind. I had ummm’d and ahhh’d about going to Business of Software for the past 2 years and always left it too late, but having been to ESWC last November I’ve change my mind about conferences and despite the long distance and travel costing more than the conference I thought I should go.

I’m glad I did. But I’m going to talk about Las Vegas and The Hoover Dam first.

Las Vegas

I live in the UK, in the flat fenland just north of Cambridge. Its an area of 1,000+ quare miles drained by Dutch engineers a couple of hundred years ago. Its not the warmest place in the world, nor the coldest. But suffice to say after the confines of the airport, then the aircraft and then standing in US border control for an hour, when you step outside (after I think 14 hours) the heat hits you like a wall. I was surprised. I knew it would be hot, but it was a shock.

Las Vegas is an incredible place of excess, bold, brash, bright lights and gargantuan statements of wealth, mainly expressed in the form of ever taller and grander buildings. Donald Trump’s building (which has an enormous terrace) demonstrated this quite well.

Due to my jet lag (8 hours out of sync) I never really got in sync during my brief stay in Nevada. As a result I’d wake up at strange times and look out of my hotel (which I initially thought was tall until I viewed Las Vegas from the other end of the strip) I’d always see people walking around the strip. The place really does not sleep. I walked from one end of the strip to the other on both sides of the road just to see what it was like. Considering the amount of alcohol being consumed (and carried around in giant plastic cocktail containers) there was very little disorder. If it had been the UK on a Saturday night it would have been a full on street fight in most cities.

In addition, I have never been anywhere where women could wear what they wanted and walk down a street unmolested. Its so warm that even at 2 in the morning I saw women dressed very provocatively safely walking down the street. Incredible.

So full marks for law and order (on the strip at least).

But those women are tourists. The women that work in Las Vegas, that is a different thing altogether – dancing on tables, strip clubs and “private” dances. And for 3 out of the 4 hours it took me to walk the strip I was incessantly pestered by Latinos trying to hand me leaflets for strip clubs and private dances “women to your room in only 20 minutes!”. I’m pretty relaxed about issues to do with sex, but I didn’t like women being treated as a commodity to be traded. Its really sleazy and something Las Vegas should get a grip of. I’ve been told it used to be worse. Have strips clubs sure, but don’t make me put up with your advertising for 3 hours, I’m trying to enjoy walking down the strip. I quickly learned that the worst job is handing out these flyers. It seems to be that only Latinos hand out flyers (which says a lot about their job prospects, really sad).

In terms of money well used to create more money I’d give Las Vegas 10/10 but in terms of money well used to improve the lives of others I think anyone creating a startup (even if it fails) is spending their money and time more wisely. Las Vegas is such an incredible demonstration of how not to use your money well. Interestingly every attendee of MicroConf that I met when they found out Las Vegas was my first visit to the USA, all said “The rest of the USA isn’t like this” in a tone that clearly indicated they also didn’t think much of Las Vegas.

I found this wonderful reproduction poster from World War II in the Victoria Pub inside the Riviera Hotel. Never seen anything like this in the UK.

It seemed that the Latin community seem to have drawn the short straw in Las Vegas and only get hired to do menial jobs. If you are white, black, asian you’re OK, but if you’re Latino, bottom of the heap. That may be an incorrect conclusion, but I didn’t see anything to change my impression of that during my short stay.

I saw very few homeless people. But of the few people I did see that were homeless some were in a real mess. I’ve never seen anyone bearfoot (except out of choice) before. I’ve since been told by people I know that have worked in Las Vegas that much of the homeless folks you do see are down-at-heel gamblers and you are just as likely to see them a few weeks later dressed in Armani. What a strange life some folk lead.

Hoover Dam

On the Sunday after my arrival I went on a coach trip to Hoover Dam. Very impressive structure.

The spillways were incredible as was the size of the generators in the turbine hall. They made the forklift trucks next to them look like matchbox cars. Getting into the tour for the turbine hall was like going through airport security all over again. There was a potted film about the construction of the Hoover dam that lasted 15 minutes (and is a pale shadow of the awesome documentary on the same subject made by the BBC – if you get a chance to view the BBC work do take it).

If you go to Hoover Dam do take the turbine hall tour, you’ll see some things you had no idea existed. Well worth it.

If I’d had a whole day spare I’d have gone on a trip to the Grand Canyon but you get picked up at 6Am for that tour and I didn’t find out about it until 11AM on my only free day.

Micro Conf

OK, so I didn’t like Las Vegas. What about MicroConf? I liked it. A lot.

I was the only attendee from Europe. My blood sugar was all over the place and I was 8 hours out of sync. As a result I really wanted to sleep at the end of each day before the meetup. The problem was on both Sunday and Monday evenings I overslept and missed much of the meetup. If this had been Europe, arriving at the meetup at 10:30PM wouldn’t be an issue – people would still be around and talking/drinking until 2 or 3 in the morning, but American culture is different and everyone faded away by about 11:30. So I didn’t make the most of those meetings, but I can’t fight my blood sugar levels and body clock.

The conference itself is a 2 day, single track conference.

The single track is one of the aspects that I really liked and that many other conference organisers would do well to note. The single track means that everyone gets to hear/see the same talk. So when you meet someone you have that in common even if the work you do is vastly different. A single track also prevents choice paralysis – which track should I go to? If I go to this I have to miss that and if I see that I have to miss that. Such a dilemma. And quite real too. Sure there may be a speaker you know nothing about or you already know their stuff and don’t want to listen to them. OK, so go for a walk, answer email, read a book, play angry Birds.

It seems that feedback indicates this went down well with many other attendees and future MicroConf (if there are any) will also be single track.

Of the attendees I did meet there was quite a range of businesses present, from software developers (surprise!) to educators to bridal shop owners to registered nurses to people selling products I didn’t know existed (Paleolithic diet food). Despite the variety of businesses only 5 of the 110 people were women. Hopefully there will be better representation next year.

The main themes that seemed to stand out were (in no particular order):

  • Testing – as in A/B testing, not unit testing. Test everything.
  • Email marketing.
  • Failures teach you as much (if not more) than success.
  • Do consulting or do your own business. Do not try to do both.
  • Read.
  • Create products with recurring revenue.
  • Automate as much as you can.
  • Web based apps preferred compared to desktop apps.
  • SSD based laptops are fast – based on chatting to MacBook Air owners.

The speakers were all approachable, but as usual when I meet someone “famous” my mind goes blank and I can’t think of the questions I had wanted to ask. So now you know, you are not alone. On Tuesday evening Andrew Warner very kindly hosted a party in his suite because we couldn’t get the Victoria Pub to ourselves. Thanks Andrew. That was a good chance to talk to some people I’d missed before and get a photo of Rob’s famous ‘create’ tattoo.

I recorded all the speakers (except one for some reason) and have written their talks up. All of the speakers but one have given me permission to post their talks as MP3 (and Rob and Mike have also said yes). I have 20 pages of notes I made during the sessions. I didn’t think that was much but Rob was impressed. I’ll be posting my take on the talks and the MP3 of each talk here during the next few days.

Would I go to MicroConf again? Yes. But I do hope its not Las Vegas next time 🙂

Perl Profiling API

By , August 6, 2011 3:48 pm

For some time we have had tools for use with Perl in the beta stage. The reason the tools never left beta is because the instrumentation part of the tools was tool brittle. It relied up scanning the Perl binary for suitable insertion points then modifying the binary to insert a suitable profiling callback. The callback would then ask the interpreter for filename, line number and function name information.

This worked fine so long as the version of Perl being used had data stored in the same format as the version we had inspected when we wrote the tools. As soon as the internal design of the Perl runtime changed our tools stopped working.

The Perl interpreter has some rudimentary profiling capabilities built in. The are only suitable for gathering timing statistics and cannot be used to provide code coverage flow tracing or call graph timing results. The only was to get statistics for these uses is to build your own profiling API. That is what we have done with the Perl Profiling API.

The Perl Profiling API is inspired by similar functionality in the Python interpreter, the Ruby interpreter and the Mozilla JavaScript interpreter. The API consists of four functions:

  • void Perl_set_do_profiling(int enable);
  • int Perl_get_do_profiling();
  • void Perl_set_coverage_callback(PERL_PROFILER_CALLBACK callback, void *userData);
  • void Perl_set_profiling_callback(PERL_PROFILER_CALLBACK callback, void *userData);



                               const char          *fileName,
                               const int            lineNumber,
                               const char          *functionName,
                               const char          *packageName,
                               void                *userData);

The PERL_PROFILER_EVENT event can receive values of PPE_LINE, PPE_CALL, PPE_RETURN. PPE_LINE is used for line visits, PPE_CALL is used when a function is called, PPE_RETURN is used when a function returns. The fileName and lineNumber specify which part of which fileName the code is executing, the functionName specifies the current function and the packageName specifies which package/namespace the function is part of (if any). The userData value is the value passed when setting the callback using Perl_set_coverage_callback() and/or Perl_set_profiling_callback.

void Perl_set_do_profiling(int enable);

Perl_set_do_profiling() is used to enable or disable profiling with either of the callbacks. Pass TRUE to enable and FALSE to disable.

int Perl_get_do_profiling();

Perl_get_do_profiling() is used to determine if profiling is enabled or not. The function returns TRUE for enabled and FALSE for disabled.

void Perl_set_coverage_callback(PERL_PROFILER_CALLBACK callback, void *userData);

Perl_set_coverage_callback() is used to set the callback that will be called for line visit events. The callback function will be called for all lines the Perl program visits. The userData value specified will be passed to the callback.

void Perl_set_profiling_callback(PERL_PROFILER_CALLBACK callback, void *userData);

Perl_set_profiling_callback() is used to set the callback that will be called for function call and function return events. The callback function will be called for all functions the Perl program visits. The userData value specified will be passed to the callback.

Using the Perl Profiling API

By default the Perl Profiling API is not enabled and no callback values are set.

To use the Perl Profiling API we need to define a callback to do the work.

void callback(PERL_PROFILER_EVENT event,
              const char          *fileName,
              const int            lineNumber,
              const char          *functionName,
              const char          *packageName,
              void                *userData)
    case PPE_LINE:
        storeLineVisitInfo(fileName, lineNumber, userData);

    case PPE_CALL:
        startFunctionCallTiming(fileName, lineNumber, functionName, packageName, userData);

    case PPE_RETURN:
        endFunctionCallTiming(fileName, lineNumber, functionName, packageName, userData);

The implementation of storeLineVisitInfo(), startFunctionCallTiming() and endFunctionCallTiming() are down to the profiler writer. You can use these to implement a timing profiler, code coverage, flow tracing or some other tool.

We also need to tell the Perl to use the Perl Profiling API.

	Perl_set_coverage_callback(callback, userData);
	Perl_set_profiling_callback(callback, userData);


As you can see using the Perl Profiling API is easy.

Adding the Perl Profiling API to the version of Perl you are using is also straightforward – there are two files that implement the API and two simple modifications to the runtime loop in dump.c and run.c. The overhead of the API is trivial when not in use and when in use the overhead is defined by what the callback writer chooses to do.

Learn more about the sources and binaries for the Perl Profiling API.

Panorama Theme by Themocracy