http://www.lysglimt.com/ Hans Lysglimt personal blog about life, travel, investing and much much more. en Copyright 2006 Thu, 14 Dec 2006 17:55:07 +0000 http://www.sixapart.com/movabletype/ http://blogs.law.harvard.edu/tech/rss More experiences from AdWords and AdSense for good measure: AdWords - It is about having an ad that fits into the mindset of the searchers. - If you need to set up a new domain and website just to fit into the mindset of the searchers. If people are searching for "49ers coffie mug", your "healthycoffie.com" domain name might just not cut it. You might need a different domain name like "custommugs.com" or "sporsmugs.com" just to get the click at all. - Small improvements in click through rates can result in huge price drops. - Test test test - you can not tell what works or not unless you test it. - Lists are a given, people love lists. AdSense - Beware of smart pricing, keep to good quality. - Smart pricing can easily lead to a race to bottom prices. - Keep a log of the changes you make, otherwise you will not be able to know what works. http://www.lysglimt.com/2006/12/adwords_adsense_best_practises.html http://www.lysglimt.com/2006/12/adwords_adsense_best_practises.html Internet Marketing Thu, 14 Dec 2006 17:55:07 +0000 http://www.lysglimt.com/contact.html]]> http://www.lysglimt.com/2006/12/updates.html http://www.lysglimt.com/2006/12/updates.html Hans Lysglimt Thu, 14 Dec 2006 17:43:32 +0000 Explorer Thor Heyerdahl once told me in a private conversation that "we are still living in the middle ages", and he was so right. I am just looking at the internet and how unsophisticated it is, it has just only started. This will happen in the next 5-10-20 years, who knows what will happen after that. - All human formal knowledge will be entered into the semantic web. - Machines will then be able to read human knowledge and to a limited extent interpret it. This will not be real artificial intelligence, but something fairly close in some instances. Like when the computer Big Blue beat Casparov in chess, machines will be able to amazing tasks. - We will all have small handheld computers in our hands that we will communicate with constantly. - Computers WILL be able to interpret human speech at an acceptable level, this is not a very sophisticated task. - The economy will then become MUCH MUCH more efficient. Supply and demand will be instantly available. Whenever you need something the suppliers will be instantly there. Machines will order what you need for you and deliver forward in the value chain. - You time will be MUCH MUCH more efficient. Machines will tell you what to do. How to do it most efficient. When you need to do a task machines will instantly suggest you the best procedure. - Machines will suggest good fits for you. Your old trusted best friend from college is going to the movies tonight, why not show up and have the machine automatically book the seat next to him? The combinations are endless. - Search engines today are incredibly primitive. It is almost ridiculous. Do a search for "hotel london", the search gives you nothing of value. Do a search for "hotel london" on Expedia and it might remember the hotel you stayed at last time - that is unsophisticated but shows us the direction we are heading in. Our search results will be MUCH MUCH more sophisticated and tailored to our history and fit to our needs. - Right now click advertising is the big thing. But the text click advertising like we see it today will not survive for long. In fact text based click advertising is a great example of how short we have come in the evolution of the internet. Of course the amazing effects of supermodels posing with the products we desire will grow into the net. 20 years from now you will not click on a text ad for a flight to Bali, you will see the current supermodel on the beach in a photo. - Email will evolve tremendously. We will get rid of the spam problem, we will have a controlled environment where messages and task/project management and teamwork will be fully integrated. If you consider all these things combined, and many other things we can not even imagine right now you see that the future ahead is very bright indeed. We are still in the middle ages, the internet, the information revolution and the lightening up of the global fiber network was just the last spark of the enlightenment that was needed to propel us out of the dark ages. http://www.lysglimt.com/2006/10/the_late_middle_ages.html http://www.lysglimt.com/2006/10/the_late_middle_ages.html Comments Sun, 15 Oct 2006 09:52:35 +0000 I am doing a lot of tweaking on Google AdWords. I spend USD 50 000 - 100 000+ per year on AdWords. Some things I have found - sharing it under the "what goes around, comes around" principle: - Anything "free" will allways give a high click through rate. - "FREE" is better than "Free", and much better than "free". Use capital letters when allowed. - First letter in domain name shown should be capitalized, www.Runbox.com not www.runbox.com. - You might even try capitalizing the first "W" and the "C" like this Www.Runbox.Com. - In display URL I have not yet decided on having a slash "/" after the URL or not. Need more data. - Try many ad variations, make new and delete non performing. - Be sure to delete the ads that are not working so well to improve the average click through rate. - Putting the price in the ad will send much fewer, but much more high quality clicks. But, the campaign will be less efficient, this is a paradox with AdWords. They will somehow fix that I ams sure, but for now a paradox you will have to work with. - Use the actual key word people search for in the ad header. Repeating it works fine. - As allowed get a call to action in like "click", "easy", "yes" etc. - Old deleted campaigns are gold, revitalize them with new content. Google cares about old stats. - Do not pay more than 15-20 cents per click. It is usually not worth it. Cap your clicks at USD 0.15 and improve them, you will see the price gradually drop after a few days. The better you make them the lower priced they get - it's amazing. - Make many campaigns and improve them individually. - Separate USD/UK/Australia from "Rest of the world". As a rule of thumb a click from the first group is worth twice as much as a click from the second (many exceptions naturally). - Be careful changing ads with good stats, you might loose the stats history. - Use negative keywords, make a list of them. Allways negative "-sex", "-porn" etc. Build your list ofer time and use it every time. I have a list of about 500 negative keywords for my campaigns, you will have to decide on "-free" etc. Suggestions for Google - you do read my blog I assume: - Have a box on every campaign "Add my negative list". Suggest some negative words in this list and have me add to a global list. - Allow suspended ads to be resubmitted as they are. - Allow to pause some ad copy instead of having deleting it. - Speed up the reporting to almost real time. - Do not assign the Google.co.uk team to European accounts automatically. The Europe team is not a good as the US team, this is a bad practice done by many US companies. - Show us more about what is going on. Why do prices change? - Show total number of searches on key words, so we can see the potential volume. - You write some best practises like this one, so we dont have to. There is way to much guessing and witch science going on with AdWords. http://www.lysglimt.com/2006/10/google_adwords_copy_best_pract.html http://www.lysglimt.com/2006/10/google_adwords_copy_best_pract.html Internet Thu, 05 Oct 2006 08:09:03 +0000 In business school you are taught to work for win-win relationships. You are taught business ethics. You are taught business law. You are fostered to behave in an idealized nice way - and many if not most buy into this. You are bottled up to be a nice guy and behave nice. But the business world is full of people that are anything but nice, there are bad guys all over the place. When you look around in the real business world you see people behave unethically, greedy even criminally and getting away with it. The famous cases like Enron, options scandals, savings and loans etc are just the top of the iceberg of cases. The fact is that businesspeople are behaving unethically and doing so because it pays, most get away with it. It is much like the priest who preaches faithfulness on Sunday and then screws around all week. He is preaching an idealized nice way that does not correspond with the actions of the real world. The thing is that the nice guys end up being the suckers. The nice guys who do behave ethically in a Boy Scout kind of way get screwed over by the bad guys. But business school never tell you this, business school made you expect people to be nice like the textbooks describe. In this way business school is failing miserably to prepare young people for the real world. The kind of knowledge needed is street smarts; this is also why many spectacularly successful businesspeople never went to business school. I once met with an American businessman. We came to talk about win-win relationships. "Yeah he said - win - win relationships, sure. Slapping his hands twice he said, win-win means you win twice you fuck your customer twice." The problem is that when nice guy "you" are dealing with this bad guy businessman he might talk about win-win, and you might buy what you believe are his true genuine good intentions. You might even trust him. When he then goes for seconds, and proceeds to f*** you for the second time - you are will be totally unprepared and vulnerable. Business schools would prepare students better by offering courses like "Street Smarts 101", "Advanced Suing", “Applied Game Theory”, "Contract Strategy" and "Social Control Mechanisms". http://www.lysglimt.com/2006/09/business_schools_are_full_of_s.html http://www.lysglimt.com/2006/09/business_schools_are_full_of_s.html Comments Fri, 22 Sep 2006 07:37:30 +0000 When you are a startup you sell the future. You should be aware of this. It is interesting to see how a supplyer looks at you. Say you can bring USD 100 000 in revenue to a supplyer. What is this worth to him? If the supplyer has a 10% profit margin, he is making USD 10 000 on you. But you are worth a lot more than this to him. When he is promoting his company to investors this is multiplied with a price/earnings of 10-15-20, perhaps even more. Most liklely your startup is considered one of the growth prospects that give high p/e. This means you as a customer is worth to him perhaps twice of what you are anually paying. You as a customer could be worth USD 200 000 to him. Turn this around, loosing you, or sending you to bankrupcy is a loss of USD 200 000 in value for him. Getting another customer paying USD 100 000 a year might cost him USD 50 000 in sales efforts. Sometimes his company will be based on price/sales, this can range from 0,5 to 1,2 or even 3 times sales. Your USD 100 000 cold revenue could be worth a hot USD 250 000 capitalized value to your supplyer. This is something to consider if you are strapped for cash, need funding, need to ask your supplyer for a break or heck you can even build a company on this principle. You as a customer is worth a lot for your supplyer, so he will most likely be willing to streatch himself far, even take risk to help you get over any temporary struggle as long as the future seems bright. Your supplyer is your friend. You are selling the future. http://www.lysglimt.com/2006/09/your_value.html http://www.lysglimt.com/2006/09/your_value.html Internet Fri, 22 Sep 2006 07:23:41 +0000 Who are clicking all the Google ads? I do not click Google Adwords ads, well perhaps I do but very very seldom. Say I click one ad every two weeks just for the argument. You are probably like me, you just do not click on ads since unfortunately they seldom lead to anything worthwile. Say I click on one every two weeks, that is 26 clicks in a year. At an average cost to the advertiser of say 30 cent that adds up to USD 7.8 woth of clicks in a year. There are about 1 billion internet users worldwide, say Google has a 30% market share of this. That is 300 million times USD 7.8 or USD 2.34 billion. Googles revenue for 2006 should be something like 6-8 billion. http://www.lysglimt.com/2006/09/whos_clicking.html http://www.lysglimt.com/2006/09/whos_clicking.html Internet Fri, 22 Sep 2006 07:13:21 +0000 Kopi av leserinnlegg på Digi.no: Den som først utvikler en velfungerende metodikk for IT outsourcing fra Norge og Norden kan skape noe stort. Vi er enige i konklusjonen fra Ethos, utvikling vil flytte ut av Norge. Logikken er klar, timeprisen i Norge er nå 1000+ kr, i Filippinene kjøper vi timer for 30 kr timen. Men, make no mistake, dette er vanskelig å få til. De aller fleste prosjekter så lang har mislykkets. Kina er spennende, vi har kjøpt utvikling derfra, men Kina har også en rekke problemer. På kort sikt er språket det viktigste, de er ikke gode i engelsk. Dernest er miljøet svært dårlig, Kina er ikke et sted å la barna vokse opp. Geopolitisk risiko skal man ikke glemme, før eller senere smeller det ref Tinanmen, Taiwan mm. Til sist har Kina en tikkende aldringsbølge om 20-30 år. Selv satser vi på Filippinene. - Gode i engelsk, det er offisielt språk i Filippinene. - Kulturmatch med Norge. Norsk sjøfart har lenge foretrukket Filippinere de er hyggelige og joviale. - Hyggelig rent miljø utenfor sentrale Manila. Blogger en del om dette på: http://www.lysglimt.com Hans Jørgen Lysglimt http://www.lysglimt.com/2006/09/outsourcing.html http://www.lysglimt.com/2006/09/outsourcing.html Outsourcing Wed, 13 Sep 2006 09:52:11 +0000 Amazing the deals you can get on last minute travel from Oslo. One week on a greek island, flight and hotel for USD 135 must be a bragain any way you look at it. Apollo Gardermoen Zakynthos Zakynthos Uspes Zakynthos Dobbeltrom. 2-3 pers. Tir 5 sep 6:15 Tir 12 sep 14:50 7 NOK 848 Apollo Gardermoen Bulgaria Burgaskysten/Bulgaria Uspes Burgaskysten Dobbeltrom. 2-3 pers. Tir 5 sep 16:10 Tir 12 sep 23:50 7 NOK 848 Apollo Gardermoen Thessaloniki Halkidiki Uspes Halkidiki Dobbeltrom. 2-3 pers. Man 4 sep 7:00 Man 11 sep 16:25 7 NOK 948 Apollo Gardermoen Preveza Lefkas/Parga/Restplasse Uspes Lefkas/Parga Dobbeltrom. 2-3 pers. Man 4 sep 17:45 Tir 12 sep 2:05 7 NOK 948 Apollo Gardermoen Zakynthos Zakynthos Uspes Zakynthos 1-romsleilighet. 2-3 pers. Tir 5 sep 6:15 Tir 12 sep 14:50 7 NOK 948 Apollo Gardermoen Djerba Djerba/Tunisia Uspes Djerba Dobbeltrom. 2-3 pers. Tir 5 sep 20:20 Tir 12 sep 19:30 7 NOK 948 Apollo Gardermoen Kavalla Thassos Uspes Thassos Dobbeltrom. 2-3 pers. Tir 12 sep 6:00 Tir 19 sep 13:55 7 NOK 948 Detur Gardermoen Tunisia Monastir Uspesifisert Reise Rom/leilighet Lør 9 sep 12:15 Lør 16 sep 11:15 7 NOK 990 http://www.lysglimt.com/2006/09/last_minute_travel.html http://www.lysglimt.com/2006/09/last_minute_travel.html Travel Sun, 03 Sep 2006 08:43:03 +0000 Easy things to outsource are: Single bug fixing Graphical design of graphical elements Simple one on one support Install software packages ### This is really Micro Outsouring. This can be done without on site precense. A bit more complicated: Webdesign of full websites Works on good software packages where the end result is clearly described Clearly identified software fixes that one person can do ### This is really Micro Outsourcing. This can be done without on site precence. Hard: Software fixes that are not clear from the outset Software development Very hard: Ongoing efforts over time where Service Level Agreements SLAs must be clearly defined. Collaborative efforts where different individuals or groups must work togheter and coordinate efforts ### To be successful at this you might need on site precence. This lifts the whole effort up to a completely new level. As previously mentioned the project management part is the hardest part of outsourcing. One way to go is to break down complicated outsourcing jobs into many easy micro outsourcing jobs. You can then also set several teams on doing the different micro jobs and build as if playing with Lego. This does demand quite a bit from the project manager. The question is: can outsourcing efforts be self managed? Can you hire hundreds or thousands of developers in the developing world at USD 1+ per hour and have them self manage to create a valuable product or service? It is a billion dollar question... http://www.lysglimt.com/2006/09/micro_outsourcing.html http://www.lysglimt.com/2006/09/micro_outsourcing.html Outsourcing Sun, 03 Sep 2006 08:26:36 +0000 There is a decent chance I will go to the Philippines this winter. A good friend of mine has set up an outsourcing operation in Paranaque in Manila. Will fly through Hong Kong. http://www.lysglimt.com/2006/09/philippines.html http://www.lysglimt.com/2006/09/philippines.html Travel Sun, 03 Sep 2006 08:23:47 +0000 www.gumball3000.com ]]> http://www.lysglimt.com/2006/08/gumball_3000.html http://www.lysglimt.com/2006/08/gumball_3000.html People Mon, 28 Aug 2006 09:52:05 +0000 I am working on developing best practices for outsourcing projects. Outsourcing provides some major challenges. The solution is to formalize the processes. This might look cold and stiff, but it will provide a framework for more human person to person contact. Instructions for outsourcing projects - Draft. Instructions For Outsourcing Projects: Quality Quality Quality In general: Welcome to project XXXXX. The main reason we do outsourcing is the cost advantage of going offshore. The cost difference is usually more than 1:10 between the offshore and onshore cost. We therefore want to outsource as much of the job as possible. Most of the cost in most outsourcing projects are actually the onshore project management cost. A major aim is therefore to reduce the project management cost. Key points that should be cleared when starting any offshoring project: - How much USD will the client cover and reimburse for miscellaneous petty cost? A small amount like USD 20-100 is usually a good idea to cover small incurring cost to buy needed licences, services etc. - State what time zone the client is in. Use this time zone in the communication. Major points: - Think long term. Build quality into everything you do. - Make it possible for someone else to continue upon and build upon your work. - Use as little of your clients time as possible. - Use standard components, standard protocols, standard interfaces. Instructions: - Plan your work well and show a clear plan for the procedure you intent to follow. - Structure your work well. - Document what you do well. In the code and in reporting. - Preferably use Open Source code. - Preferably use Open Source components that are live and supported. - Research the software components you would like to use and compare them with alternatives. Show why you have chosen to use one component over another. - Write as little custom code and custom solutions as possible. - If you need outside help for something. Suggest who and how this will be done. - Provide links to the Wikipedia article whenever you introduce concepts that might be unfamiliar to your client. - Provide examples with URLs. - Show don’t tell how you suggest it. You are not allowed to represent client XXX in any way. Do not contact any third party without prior agreement with you client. - Work to use as little of your clients time as possible. Try to answer any questions yourself. Ask around in forums etc for answers instead of asking your client. Suggest solutions to your client. Have alternatives when suggesting things. - Do not ask your client to do something you can do yourself. Example: If you need a webhosting account to upload a sample, find it and set it up yourself, do not ask your client to "get you a webhosting account". - At last: Before you send off an email, status report etc. Have you gone over these communication instructions? Did you follow all the instructions? - If you see that the project will be delayed communicate this immediately. - Clearly state at the end of the communication: # Are we on target # Who is working on this project, provide the names of the individuals. Do not assume that someone is good enough, always clearly state who and how much time. # How much time are they devoting to this project # Will there be delays # What is the next action to be taken and by who: Preferably state a clear progress that you will make without further instructions from the client. Examples: "I am working full time on this now. I will research these four suggested alternatives and report back to you within four days." "I and the two new Java coders (Andrew Giilupra and Tony Perez) will be working full time both Saturday and Sunday here in the office and will send a workable prototype back to you Monday morning your time" "I am awaiting your answer." (Not a very good message but sometimes necessary of course.) http://www.lysglimt.com/2006/08/outsourcing_instructions_draft.html http://www.lysglimt.com/2006/08/outsourcing_instructions_draft.html Outsourcing Sat, 26 Aug 2006 12:32:07 +0000 I am working extensively on outsourcing. As the cost per hour of developers in Norway is now approaching USD 100-200 we are forced to go offshore. I have had good experience with offshoring to India, China and the Philippines at USD 2-10 per hour, the logic is more than obvious. What are the challenges of outsourcing then? The #1 challenge is project management. I have found that for every hour I buy offshore we need to spend another quarter to half hour in project management in Norway. Reducing the cost of project management in offshoring is one of my major priorities right now. http://www.lysglimt.com/2006/08/outsorucing.html http://www.lysglimt.com/2006/08/outsorucing.html Outsourcing Thu, 24 Aug 2006 12:01:51 +0000 http://www.linkedin.com]]> http://www.lysglimt.com/2006/08/linkedin.html http://www.lysglimt.com/2006/08/linkedin.html People Mon, 21 Aug 2006 22:34:01 +0000