No blog is complete without regular mentions of twitter. If you still don’t ‘get it’ watch this video. (via Dan Hite of spellquizzer.com)
Digital River up to their old tricks?
Popular blog BoingBoing have reported on esellerate trying to sneak worthless and unwanted extras into their shopping cart. Some of the commenters on the post seem to think this could be an honest programming mistake. Given that esellerate’s parent Digital River is widely known and reviled for dubious upsells, I doubt it. Any vendor that chooses to use a Digital River subsidiary for their ecommerce has only themself to blame if Digital River does the same to their customers. Have you checked your shopping cart recently?
The two types of reseller
In theory, the Internet allows customers to find products without the need for middlemen (unless you count Google). In reality, the age of full disintermediation has not yet arrived, and perhaps never will. Middlemen are still important. One of the more important types of middlemen for a software vendor is the reseller. However it is important to realise that there are two completely different types of reseller. One can be very useful to you as a software vendor, the other is generally a pain in the backside. They should be treated accordingly.
Value added resellers
A value added reseller buys your product and then resells it to their customers. Usually they will buy from you at a discount and resell to their customer at full price, pocketing the difference. Typically a reseller will expect between 25-50% discount depending on a range of factors including what services they provide (e.g. localisation and support), the price of the product and volumes sold. Often the discounts are on a sliding scale, with the discount increasing with sales volume. This type of reseller can add value for you and the cusomer in various ways. They may:
- have expertise in markets that you don’t
- be able to reach markets that you can’t, due to barriers of language, culture or geography
- localise your product and marketing materials
- provide first level support.
- sell your software as part of a package including other services, software and/or hardware.
However there are a few things you have to look out for:
- You don’t want to end up competing directly against your own resellers in existing markets.
- Resellers may undercut you on price.
- Beware of offering any sort of exclusivity.
- Customers may sign up as resellers just to get a discount for a single purchase.
You can try to avoid direct competition with resellers by only picking resellers in different markets. You may also be able to specify certain terms and conditions, for example that they can’t bid against you on certain phrases in Google Adwords. But, depending on the law where you live, you can’t tell a reseller what to charge the customer. For example, under UK law, this is considered ‘price fixing’ and is illegal. So make sure you charge a reseller a percentage of your recommended price, not a percentage of their sale price, to make price cutting less attractive. For example, if your product retails for $50, charge the retailer X% of $50. Not X% of what they sell for. Otherwise they could undercut you by selling for $30, and still make a profit, or even give your software away for free.
Resellers will often ask for exclusivity. Exclusivity can provide extra motivation to the reseller (a reseller won’t want to put a lot of effort into marketing your software if you can pull the plug at any time for no reason), but what if the reseller loses interest in your product? It happens. You could be left in a very bad position unless you can terminate the agreement. So you should agree some sort of minimum volume of sales for a reseller to retain exclusivity.
Offering a sliding scale discount with no discount for the first purchase will discourage customers from trying to take advantage of reseller discounts.
Value subtracted resellers
The other type of middleman that call themself a reseller are really just acting as outsourced purchasing for your customer. They buy your software on the customer’s behalf so that the poor darlings in the customer’s accounts department only get a single invoice for software sales per month, instead of one per vendor. They don’t add any value at all as far the vendor is concerned. In fact they just make the vendor’s life more difficult by getting between the vendor and the real customer. I had one reseller of this type who, after some twenty emails exchanged, failed to workout how to buy my software from my website. How dim can you be that you can’t click a ‘buy now’ button and fill in a few details when that is what you do for a living? Hence I call them ‘value subtracted resellers’.
This type of reseller will often ask for a discount. Don’t give it to them. The real customer has probably already instructed them to buy your product, so a discount won’t help to close the sale. Also the reseller might pocket the discount instead of passing it on to the customer. If anything, charge them more.
GoogleCheckout price increase
It was always on the cards that Google was going to raise the prices of their payment processing service, GoogleCheckout. Up till now I had effectively used GoogleCheckout for free, as they offset their 1.5% + £0.15 processing fee against my Adwords spending. But they are dropping the Adwords offsetting and introducing a new tiered pricing structure.
As I put most of my payments through PayPal I will probably be on the highest price tier of 3.4% + £0.20 per transaction. This means that a £19.95 sale will cost me £0.88 (4.4%) through GoogleCheckout as opposed to the £0.68 (3.4%) I pay through PayPal. I wouldn’t mind the higher fees if I got a better service than PayPal. Unfortunately GoogleCheckout still has all the flaws I commented on back in April 2007, namely:
- Google still don’t accept payments in more than one currency (e.g. as a UK resident I can only accept payments in £). Expecting anyone outside the UK to pay in £ is a very bad idea.
- Delays between customer purchase and payment clearance result in angry and/or confused emails from customers wondering why their licence key hasn’t arrived. This has improved since the early days of GoogleCheckout, but it is still an issue.
- Google’s option to anonymise the customer email address is a royal pain in the backside for the vendor. It causes me of lots of wasted time and unecessary emails.
- The customer *has* to sign up for a GoogleCheckout account (unlike PayPal).
- There is a £7 chargeback fee (PayPal don’t charge a chargeback fee).
- Customers prefer PayPal.
About the only advantage of GoogleCheckout is the GoogleCheckout badge that appears alongside your Google Adwords ad. In their email to me explaining the price rise Google claim:
Advertisers who use Checkout have the opportunity to display the Checkout badge on their ads, which has proven to be an effective way to differentiate ads and attract user interest. Checkout users click on ads 10% more when the ad displays the Checkout badge and convert 40% more than shoppers who have not used Checkout in the past.
My own measurements showed a worthwhile effect from the GoogleCheckout badge, but I am not convinced it is worth the additional problems and expense of GoogleCheckout just to get the badge.
I already push PayPal more than GoogleCheckout (e.g. you have to click a link from my US dollar payment page to see a GoogleCheckout button). The price increases will probably result in GoogleCheckout being pushed further into the background for use just as an alternative for those that don’t like PayPal. I don’t know if Google will punish me by removing my Adwords badge.
Note, in order to continue to use GoogleCheckout from 5 May 2009 onwards, you must login to your account and accept the new Terms of Service between 18 March and 4 May.
T-shirt update
I’ve have just made the first payout of royalties from T-shirt sales. $106.20 of Zazzle royalties were split evenly between Sightsavers and JaipurFoot. Patrick McKenzie has also made a very generous additional donation as he promised on his blog.
Sales dropped off rapidly after Xmas, so I am probably going to leave programmer-tshirts.com ticking over until Xmas 2009. Thanks again to everyone that bought T-shirts or helped with the publicity. Special thanks to Patrick for setting up the programmer-tshirts.com site on his server.
The history of marketing, in 3 minutes
The history of marketing, by Scholz & Friends (via the quirky, eclectic and always interesting DarkRoastedBlend.com ). Sound optional.
Outsourcing artwork through 99designs.com
Most software developers have horrible graphics design skills. I am no exception. So I decided to run a contest on 99designs.com for some artwork I wanted to use to promote Perfect Table Plan. I have used DesignOutput.com once before, but I decided to give their competitor a try after seeing the stackoverflow.com and homedocumentmanager.com logos created through 99designs.
I put up $300 prize money for a humorous cartoon of a seating arrangement gone wrong. Fees cost me an additional $69. The detailed brief is here and I also supplied this rough sketch:

Here is the winner I chose:

I ran it as a ‘guaranteed’ competition to try and attract more designers I also made an effort to give plenty of feedback on the designs submitted and I let it run the full 7 days. I found the 99designs website easy and intuitive to use. Setting up the competition only took ten minutes or so. My only gripe is that the ‘eliminate’ and ‘choose winner’ buttons are so close together that it would be easy to click the wrong one. You do have to confirm the winning choice however.
Most of the other competitions are for logos and websites, so mine was a little unusual. But I still got over 40 different designs (some variations on a theme) from 20 different designers. You can see the designs here (those that haven’t been withdrawn). The quality of the entries varied, but much of it was really excellent. I am very happy with the winning entry.
99designs seems rather brutal for the designers. 40+ entries is common and some contests get over 1,000 designs entered. This means stiff competition and the designers don’t get anything unless they win. I am guessing that they are mostly students, who are happy for the practice, or living in cheaper parts of the world, where $300 is a significant amount of money. But it certainly offers excellent value for money for those running contests.
WithoutAFather.com
WithoutAFather.com is a website aimed at is a website aimed at helping teens who are growing up without a father. It provides them with advice and online mentoring on various subjects, including work and business. I think it is a very worthy cause, so I am giving it a mention here.
You can sign up as a mentor to provide advice online. Currently there are only two mentors with a software background. You might need to be able to speak ‘text’ if you sign up though.
You can read more about this project in this interview with founder Sam Berns at the FollowSteph blog.
Interview with Craig Peterson of Beyond Compare

I am a fan of file and folder comparison utility Beyond Compare from Scooter Software. It is a very polished and powerful piece of software with a big following. But I was intrigued by some of the unusual decisions they had taken: competing in a market with lots of free alternatives; going 6 years between major upgrades; re-writing from scratch; releasing a Linux version; and having an extremely generous trial policy. How had they succeeded despite ignoring much of the conventional wisdom? Craig Peterson of Scooter Software kindly agreed to answer my questions.
Can you tell me a bit about your background before working on Beyond Compare.
I started at Scooter Software straight out of college. I did a lot of programming for fun before then, but this was my first professional job and my first introduction to Delphi.
How long have you been working on Beyond Compare and what is your role?
I’m the lead developer and I started here in late 2000, a few months before significant development on v2 started. Most of my time is spent working on the directory comparison and the virtual filesystem layer (ftp, zips, version control), but there are very few places in the program that I haven’t worked on. My non-development tasks include managing the build process, interfacing with our component vendors, keeping track of any interns, and tech support, when there’s difficult questions or when the dedicated staff don’t have time.
How many other people work on Beyond Compare on the business side and on the development side?
There’s one other full-time developer, one part time developer, two in tech support, two in sales, and our president, who handles everything else.
There are lots of different file comparison tools. How do you manage to run a successful business with so many competitors, many of them free?
Competing with the free tools hasn’t been as hard as I expected. The big advantage we have here is that people are paying us for the software, so we have strong incentive to provide good tech support and to provide the features they want. We work 8 hours a day on it, which gives us more time to develop new features than someone doing it as a hobby, and we can afford commercial libraries that someone providing a free utility can’t.
As for the commercial competitors, it’s mostly a matter of providing something that they don’t. In our case BC’s directory comparison is much more powerful than the alternatives, and we have viewers for other file types like images, binary files, and data files. That allowed us to keep competing even when we were lagging in other areas.
What are the main methods you use to promote Beyond Compare?
We rely almost entirely on word of mouth. We’ve had lots of customers tell us that they brought BC with them when they switched jobs and ended up getting their companies to spring for larger licenses. We do spend some money on Google AdWords, and we hired a company to periodically submit our site to search engines and download sites, but we’ve never run a print or banner ad.
scootersoftware.com has an enviable Google page rank of 6 and ranks second for “file comparison software”. Have you spent a lot of effort on SEO?
We had a company help with SEO for a lot of the 2.x lifetime, and they’d suggest tweaks to improve things. For v3 we took a different approach and redesigned the site to make it more accessible to potential customers. I’m sure some of those changes helped, since we ended up adding more descriptions and using more synonyms, but it was primarily a case of asking how we could make it easier to find the information someone would want and expanding on that. We still have a fairly wordy page title though, and I think that’s entirely for the search engines’ benefit.
Are most of your customers programmers, or does the software appeal to a wider audience?
I’d say more than 50% are programmers, but there’s definitely a wider audience. System administrators use it for migrating servers, web developers use it instead of a traditional FTP client, and non-techs use it for backups or synching their laptops and desktops.
How long do you allow people to use the trial version?
The trial is for 30 days, but it only counts days you actually run it, so if you use it infrequently you could easily go six months or more before it times out.
I used Beyond Compare for ages before I bought a licence. I would have bought it sooner if you had been less generous with the trial. Why did you go for such a long trial period?
That goes back to competing with all the other products out there. If someone installs two programs to evaluate, and then doesn’t have a chance to really try them out until a month later, the one that works is more likely to get the sale. It also makes it more likely that potential customers will learn the application and start relying on it, so when it does come time to pay they’re less likely to throw out that investment and switch to another tool.
I understand Beyond Compare v3 is a complete re-write. Why did you feel a complete re-write was necessary? Was it a good business decision in hindsight?
There were two reasons why we felt a re-write was justified: (1) we had a lot of features we wanted to implement that wouldn’t work in the current framework, and (2) we over-estimated the speed that we could re-write it.
In the text compare we wanted inline editing with dynamic re-comparisons and 3-way merge, neither of which would have been easy to integrate into the v2 codebase. The directory compare had similarly major changes, though a lot of that is internal and in preparation for other features, so it isn’t as visible. There was also all the work we did to get a Linux release out. It wasn’t a complete re-write though. Anything that didn’t need significant changes, like our reporting engine, was brought over mostly as is.
I think it was a good business decision in that it allowed us to rethink and rework a lot of things without the old baggage, but it was bad in that it significantly limited what we could release in the meantime. We ended up going 6 years between major versions and even though we were always busy adding new features, we couldn’t release them until we got back to feature parity with v2.
Beyond Compare is written is Delphi. What would you say are the advantages and disadvantages of Delphi compared to other development ‘stacks’?
I think Delphi is still the best tool for developing a native Windows application quickly. It’s very easy to mock up interfaces and then fill them in with code. The resulting exes don’t have any external dependencies, which makes redistributing them easy. The VCL (UI framework) ships with source code, and that has permeated the community, so the vast majority of third-party components also include source.
On the flip side, it still can’t produce 64-bit executables and isn’t cross platform. It doesn’t have a garbage collector or as large of a class library as Java or C#. The community isn’t as large either, so there’s usually only a couple of choices when you’re looking for specific components, and if that vendor stops developing it, there aren’t as many people to pick up the slack.
If you had to do it all again, starting now, would you still choose Delphi?
Probably, but I would seriously consider C# or Qt. In our case we have a lot of experience with Delphi and we know the libraries, so starting from scratch in another language would be a significant barrier.
Beyond Compare has a very slick user interface. Did you do any usability testing?
Not as such. Our primary usability improvements come from using our own product. We use BC every day, for every comparison we do, so anything that sticks out tends to get squashed quickly. We also get alpha/beta versions into customers’ hands as soon as possible, and keep iterating until they’re happy. V2 and v3 both had private beta tests that lasted over a year and a half, and some of the features changed dramatically in response to that feedback.
Beyond Compare has a nice integration with Windows Explorer. Was that difficult to do? Were you able to do it in Delphi?
The first version of it was submitted by a user, so it wasn’t difficult at all. It has taken a lot of refinement to get perfect though, and we were changing it all the way through to the final 3.0 release. It is written in Delphi, though I did rewrite it in C++ in order to get a 64-bit version working. FreePascal has since started producing 64-bit binaries, so we’re back to the single Delphi version again.
You have Windows and Linux versions. What did you use to write the Linux version?
We’re using a heavily customized version of Borland’s now-dead Kylix product, which was Delphi for Linux. It does allow us to compile both versions from the same source code, but it’s showing its age and I wouldn’t recommend it to anyone else. Our driving goal is to have the best Windows version we can, which means sticking with Delphi and using what’s available to produce the Linux version. If OS X and Linux support outweighed that we would use a cross-platform option like Qt or Java, but we believe the Windows version would suffer in that case.
Do you sell many licences for Linux?
We sell enough to fund its development, so it’s a successful product. It does introduce new challenges though, both in development and tech support, so if we were a smaller developer it probably wouldn’t be worth the overhead.
Do you think there will there ever be a web version of Beyond Compare?
I can see it as a possibility, and it would be interesting to explore, but it’s not something we’re looking into right now. I think it would have a different audience than the current product, and would probably never be as powerful as what we can do locally.
How did you choose the price?
Our $30 standard edition is about the same price as our commercial competitors, and seems to be the standard shareware utility price. The pro edition was priced based on our competitors, what our customers were telling us they’d pay, and what we felt the downward pressure of the freeware/opensource alternatives introduced. We keep the price low in order to make our profits on larger quantities sold, instead of a larger margin per-unit. We have increased our multi-user pricing considerably over the years though; the discounts were very steep in v1 and v2, and the feedback we got was that it was just too cheap for what it provided.
I see some translator credits on the website. Is v3 available in languages other than English?
Not officially, but we have just released beta versions of a couple of languages. The current credits are for the v2 translators, who are generally the same people working on v3 translations.
How important are resellers to your sales?
We get a lot of sales through resellers, but it’s generally from people who would buy it either way. Foreign resellers are a help to the customers though, because they allow them to order in their own language using the local currency.
Does Scooter Software have any other products besides Beyond Compare?
No, BC keeps us plenty busy.
Thank you Craig.
You can download a free trial of Beyond Compare from the Scooter Software website. I have no affiliation with Scooter Software beyond being a paying customer.
Marketing your software through affiliates
The idea of paying someone for sending you business has been around for a long time. Affiliate marketing is just a new, Internet-based, take on it. An affiliate sends traffic to your website and is paid a commission on each sale. For software this commission will typically be in the range 20-50% of the sale price (although commissions of 75% or more aren’t unheard of). Commission is usually calculated by using cookies to track the number of successful sales (‘conversions’) due to each affiliate.
In theory you can set up your own affiliate commission tracking system, but affiliates would have to trust that your system is honest. It would also involve quite a bit of wheel reinventing. Consequently most vendors use affiliate marketing systems administered by third parties such as shareasale, clickbank or commission junction. Payment processors, such as Avangate and e-junkie, also have their own affiliate marketing systems.
It sounds great. The affiliates are doing marketing for you and you only pay them when you make a sale. How can you possibly lose? In fact there are quite a few potential downsides:
- You may end up paying commission on sales through affiliates that would otherwise have come to you direct.
- Affiliates won’t be happy if there is any way to purchase where they don’t get their commission (‘leaks’). This might mean you may not be able to offer some forms of payment, such as cheque or credit card over the phone.
- You will be competing against your affiliates for search engine ranking.
- Somebody who wants to buy several copies of your software could sign up as an affiliate to get the commission on their own purchase. You then have to pay commission, but get no additional sales.
- Affiliates may compete against you in PPC ads, driving up the cost of your ads.
- Even though you make less on the sale, you still have the full cost of supporting the customer.
- Some affiliates operate at the shadier end of the market and may resort to various dodgy, or even criminal, practises to get their commission, including:
- Spam.
- Annoying pop-up ads.
- Cookie stuffing.
- Misrepresenting your product.
- Adware.
- Buying your product with a stolen credit card.
You maybe be able to prevent some of the above abuses with appropriate terms and conditions. ( I should also point out, in the interest of fairness, that there are various ways that the affiliate might lose out on commission that is rightfully theirs. For example customers who block cookies and even fraud by vendors.)
Drawing up agreements, recruiting affiliates, providing them with marketing materials, doing the accounts and paying your affiliates all takes time. You can automate quite a lot of it, but it still takes time to set-up the system, answer questions, keep everything running smoothly and check that affiliates are behaving themselves. Time that you could be spending doing other more lucrative and interesting things. As always, there is no such thing as a free lunch.
Joel Spolsky has been an outspoken critic of affiliate marketing:
We did an affiliate program, and found it to be a big waste of time. It generated only a trickle of sales; most of the people in the affiliate program would have linked to us anyway; probably 80% of the affiliates just became affiliates to get a kickback on the one item they bought for themselves or their job.
Affiliate links only works well for mega retail sites like Amazon, where an affiliate has a chance of making a reasonable amount of money.
Our affiliate program was one of those cases where we learned that time spent improving our product pays off many times as much as time spent dinking around with so-called clever marketing schemes.
Don’t waste your time. Move on. Do something to make your product better and Just Say No.
I am inclined to agree with Joel. Hard data isn’t easy to come by. But, from reading around and talking to other vendors, it seems that very few of them are getting more than 5% of their sales through affiliates. I did have a home-rolled affiliate program for Perfect Table Plan, but I shut it down because the number of sales just wasn’t worth the administration overhead. Some of the affiliates never sold a single licence. I might be more successful if I used a more automated affiliate marketing system and put more effort into recruiting higher calibre affiliates, but I still don’t think it would be the best use of my time.
I have heard that there are ‘super affiliates’ with mythical powers to drive serious sales. But these people, if they really exist, get to pick and choose amongst thousands of products to market. They are going to pick mass-market products with a proven track record and they are going to want a big commission percentage. And how do you tell who is a super affiliate and who is a wannabe? They all talk a good game.
Affiliate marketing is big business, with estimated sales of over £2 billion in the UK alone in 2006. But I suspect a lot of it is from selling ‘get rich quick’ schemes, gambling and porn – not software. Particularly not software from small companies and microISVs. Obviously a lot depends on your product and market. Perhaps if you are selling mass market software (e.g. back-up utilites or virus scanners) and you have dedicated marketing staff, it might be worth your while to run an affiliate program. But make sure you automate as much of the system as possible and be realistic about the results.
