Written by Mihai Dragan on Thursday, February 25th, 2010 ( One response )
Tags: economics, entrepreneurship, online startups, Running the business, studies
Online startups are interesting micro-economies. Most of the systems are based on several variables like number of users, conversion rate and others, directly depending on the type of business systems they do (or do not) develop.
I have been studying the whole concept and focused on some of the variables that are usually left out.
The assumptions
The online startup I’ve studied to get a glimpse into this was an hypotetical online game. I’ve assumed the the game will be launched with a PR/advertising campaign, delivering an initial traffic to the website.
The game would sell power-ups to gain revenue. Other assumptions I have based my study on are that users would enter the website, register if they feel like, drop out and erase their account if it is just not their cup of tea, invite friends which would might or might not respond and of course, play.
The conventions
In order to test the user’s behaviour we will use several conventions, following Registration Rate (RR), Spread Rate ( SR – number of invited users / month / user ), Response Rate ( ReR – number of invited users actually visiting the website / month ), Dropout Rate ( DR – users cancelling / abandoning their ), Conversion Rate ( number of registered users buying upgrades), Medium cost per product (MCP).
I have also thought of a formula to determine purchase intent (PI), directly proportional to monthly visitors increase, RR and Influence rate of registered users (SR/ReR).
The Purchase Intent is Inversely Proportional to Dropout Rate (DR) and Medium Cost per Product (MCP).
The results
Although the concept hasn’t been tested on real life scenarios (I am looking for entrepreneurs willing to share some data for the study – 100% confidential) I guess I can outline some conclusions:
- Number of initial users is important but not the most important
- Spread rate (SR) is the single most important Indicator to be taken into account. Incentivize potential users to spread news and you have a booming business. Fail to do that and all the money in the world spent on advertising won’t guarantee safe revenue returns
- The response rate (ReR) is not actually that important. It doesn’t actually matter users being influential but rather willing to share information with their microcomunities
- Conversion rate (CR) is important but it will take more than that to have a really successful product. A tenfold increase in conversion rate is useless unless the registrants numbers goes up.
- Micropayments are the best choice. Increasing Medium Cost per Product (MCP), while having a steady increase in user base, means decreasing potential revenue. Increasing cost by tenfold will decrease potential sales by more than 50%
- Dropout rate (DR) is basically irrelevant as long as the dropouts are outpaced by new registrants.
- So, if it is something you should be focusing on as an online entrepreneur this is Spread Rate, Micropayments, and number of initial visitors.
I hope this helps.
Written by Mihai Dragan on Wednesday, August 12th, 2009 ( 3 responses )
Tags: Analytics, guesswork, internet, online statistics, samples, statistics
Today I’ve received news of another statistic study. It stated 2 percent of romanians wish to relocate in rural areas. The study was conducted on a “representative sample” of 1100 people.
I will not talk about romanians relocating to rural areas. It’s boring. I will talk about the concept of statistics and the changes that need to be taken to reach the next level.
The “representative sample” of 1100 people is no more than 0.005% of the total population. So … how could this sample be “representative”?
The answer lies in the phisical ability to conduct the study. There is only that much a company can do. Back in the day, when Statistics was born as a science there was no way anyone can theoritize the appearence of something like the internet, the impact it had or the fast spread it gained.
I grew up learning on the internet. I failed once the Statistics class, not being one of my favourites in college. However, what I did learn on the internet is that samples can be way larger than the standard 1000, data can be collected easier and faster than by just “asking for answers” and many others.
Statistics need an upgrade. Because now it’s more guesswork than science.
Written by Mihai Dragan on Thursday, November 6th, 2008 ( 3 responses )
Tags: campaign management, creative, interactive campaigns, viral, viral campaigns
Viral is one of those over-hyped web2.0 buzzwords.
You want a client’s approval on the budget? Just assure him it’s going to be “viral”.
You want funding for your internet start-up? Add the word “viral” in the subtitle.
You’ve lost that Webby award? But why? Your campaign was sooo viral.
However, viral campaigns actually work and engage users. Inspired by Brendan’s post on the Viral Effect, I thought I should post some of the things I know about making “it” go viral.
Target
Making something go viral without a purpose is just an waste of resources. Figure out what you want to accomplish with your viral campaign. Write it down. Find a target and sketch a plan to get to your target.
Concept
After you have the target and you know where you’re heading, assemble the creative team, brief them on the subject and start brainstorming. Write down everything you can come up with. Nothing is stupid, everything is doable.
After the brainstorm delete the least interesting/doable ideas. Now you have a concept.
Implementation
Perfection lies within details. Plan your campaign carefully. Assign the best people to the project. After you have laid out the plan start doing it. Be flexible. Ideas may come up along the way. Keep focused but open to improvements.
Testing
After your viral campaign is ready to go be sure you test it. You wouldn’t want your funky flash game dead after it reaches 10 000 users.
Seeding
Viral effects don’t just happen. People make it happen. Agencies make it happen. Be sure to target a certain message delivery channel (be it a video on YouTube, minisite, flash game or just a picture) and stick to it. You don’t want your users’ attention distributed in dozens of places.
Talk to the influential bloggers. Ask them to write about your campaign (first of all make sure it is interesting enough to write about). Try to get social media power users to mention it. Do this and more (this is where the magic actually happens).
Management
Viral effects are like hurricanes. Once they start you don’t actually know where they’re heading. The best you can do is watch carefully and correct the campaign when needed.
Analyze and report
So you made it viral. But how viral? Track the answers, the number of users that got caught up in the viral effect, blog mentions, twitter mentions and others. Find out how many of the users took the path you’ve suggested. Report it to the client. Understand what happened and do it again. Better and smarter.
Written by Mihai Dragan on Wednesday, October 22nd, 2008 ( 3 responses )
Tags: bounce rate, kissfm, websites
It’s unbelievable what some good design and some great branding can do. One of the websites we’ve just launched has an 1.13 % bounce rate.
I know, I couldn’t believe it either but the stats don’t lie. What could cause that?
