Home > Ramblings > How to sell software

How to sell software

October 4th, 2012

In a milestone of sorts, my photo deblurring software Blurity received its first negative review today. This, after receiving numerous positive reviews on much larger sites over the past month.

While I could pick apart the author’s review, I’ll focus on two bigger ideas instead: that selling software is about managing expectations, and that selling software does not mean giving it away.

Selling software is about managing expectations.  No real software is perfect.  Real software crashes, has usability issues, and has limits in its capabilities.  Unfortunately, when users lack details about the software or don’t understand them, the expectations about the software can break free and ride well ahead of reality.  The challenge is how to keep users from being disappointed.

With Blurity, the comparison is inevitably made to the deblurring software that will “surely” be in “the next version” of Photoshop.  While Adobe did put together an impressive research demo of single-image deblurring in the autumn of 2011, they have since been mum about including the technology in commercial versions of Photoshop.  Why?  I suspect that they set expectations too high.

The most impressive images from that demo were synthetically blurred (a much easier problem than natural blur), and the deblurring results they displayed were the result of close user direction of the algorithm.  The real software, operating on real blurred images, in the hands of real users, probably can’t perform at the level of the technical demo.

The trade-off is always between perfection and shipping something at all.  Engineers and artists are notoriously reluctant to let go of anything that is not “perfect,” which is why so many products seem to die due to failure to ship.  In the case of Blurity, the decision was between shipping something that doesn’t quite meet the “Enhance!!!” Hollywood-style expectations of the populace, or shipping nothing at all.  I chose the former.

So how is a software developer supposed to keep expectations low while still running an effective sales campaign?  One way is to show numerous real-world examples on the web site, like I do with Blurity.  All of the examples are available for download in blurry and processed forms, so the user can do their own before-and-after comparison.  Going a step further, every image on the Features page also shows the exact settings used to produce the results.

Another way is to outline the limitations of the software.  This is done in the Blurity user manual.  It is also done with clear alerts display at runtime when the user attempts to do something that is not allowed.

Unfortunately, users don’t read.  I’ll admit it: I don’t always read the manuals myself.  So, I deal with this by providing great, personalized customer service.  Often times, the questions asked have already been answered in the user manual, but instead of giving a gruff “Read the manual!” reply, I simply give them the answer itself.  Quicker resolutions, happier customers.

Selling software does not mean giving it away.  If software is your business, it is there to make money. The most difficult and important change I made in the development of Blurity was to stop producing a free-trial version, and instead switch to a watermarked demo.

From April through the end of June, Blurity was distributed with a 30-day free trial. During those 30 days, it was completely unrestricted. Hundreds of people installed it, and almost nobody bought it.

I was crushed. I thought that the problem was that people simply didn’t like Blurity.

However, further investigation showed that wasn’t the case. Instead, many of the people installing Blurity had just a photo or two that they wanted to fix, and since the free trial fully met their needs, they had no reason to purchase a copy. What’s more, by the time the free trial expired, a month had elapsed, and Blurity had passed to the dark recesses of their minds — or out-of-mind completely.

My inner businessman silenced my inner engineer, and I got rid of the free trial period. Moreover, I made the demo watermark so obtrusive that there was no practical way to avoid it.

Example of Blurity watermarking

Switching from a free trial to a crippled demo was the best decision I’ve made in the history of Blurity. In the first month alone, sales increased by 2400%. Let me repeat that, just to drive the point: changing to a demo model, in which the only limitation of the unregistered program was to have a watermark on the output image, increased month-to-month sales by a factor of 24. If that was the wrong decision, then I don’t want to be right.

The heavy watermarking has been the most common complaint in reviews, and that’s fine with me.  Enough of the image is visible for the software to prove it works, and enough of the image is degraded by the watermark to ensure that there are no freebies.

The second most common review complaint? The price. Blurity used to be far less expensive than it is now, and raising the price substantially was the second-best Blurity decision I’ve made.  Blurity is priced on value, and there is nothing else on the market that can do what it can do.  While axing the free trial boosted sales volume, it was raising the price that made Blurity a viable business.

So how has that worked out for Blurity?

Up and to the right!



  1. Vince
    October 5th, 2012 at 08:49 | #1

    Glad to see the years of hard work and perseverance through turmoil are paying off. Also, congratulations on managing the inner engineer/businessman dichotomy!

  2. keacher
    October 7th, 2012 at 02:59 | #2

    @Vince Thanks! Not out of the woods yet, but at least some light is now visible. 🙂

  3. xfs@HN
    October 21st, 2012 at 22:43 | #3

    “Selling software is about managing expectations.” That’s right. What is the user’s expectation when trying out your demo? I guess they are annoyed by previous bad-looking image “quality” and looking for visually appealing results, not theoretically optimal results. I have some thoughts about the negative review.

    For poor results, how does the user feel if the results are hidden vs. shown? Here comes A/B testing blahbla.

    The watermark. The user is complaining about visually unappealing watermark destroying his feel “but the image quality being completely destroyed by the watermark”. I guess there are more than one methods of watermarking, among which there are ones less visually obtrusive. For example, I think watermark in Google Maps satellite images is one of those. Actually now I can easily think of some more methods of choosing particular areas of less details for watermarking to avoid damage user perception.

  4. keacher
    October 21st, 2012 at 23:14 | #4

    @xfs@HN I think you’re right: in sales, customer perception often matters more than optimal results. In some cases, deliberately degrading results (which Blurity does not do) can improve perception. For example, adding a bit of uniformly distributed luminance noise to an image will increase the perceived level of detail, up to a point.

    The watermark is tricky. Somewhat counterintuitively, I want the watermark to be as annoying as possible without making users angry. It’s amazing how much watermarking users will put up with. Back when I did photography professionally, I was astonished by how satisfied users were with watermarked proofs, to the point that they didn’t buy the unwatermarked versions: they simply posted the watermarked proofs on Facebook, internet forums, etc. I don’t want there to be any realistic chance that somebody would want to distribute in any way a photo processed by the unregistered version of Blurity.

  5. xfs@HN
    October 21st, 2012 at 23:48 | #5

    Well thought out about professional photography. I didn’t realize annoying watermark can turn trial users into buyers. But hey, watermark can also be advertising. I got to know Getty Images through watermark.

    By the way, what’s up with the pro version? I just found and subscribed to /download/pro_version through a site search, but this is not mentioned anywhere else. Maybe I should use the support form, but putting it here doesn’t harm. So, I intend to buy one of pro version. Should I pre-order or wait to upgrade from normal version or something else? You see where I come from, and I definitely appreciate the value of turning academic state-of-the-art into product.

  6. keacher
    October 22nd, 2012 at 00:15 | #6

    @xfs@HN It seems you’ve found one of my A/B tests. 🙂 I intend to release the pro version eventually (hopefully, soon), but I’ve been testing various ideas related to its promotion. Not all users are told about the pro version; you appear to have been part of the control group prior to your site search.

    Regardless, thanks for your interest! I look forward to being able to announce the availability of the pro version.

Comments are closed.