Software Trial Periods

There’s a conversation at Contextures regarding Tableau software. One aspect of that conversation, the trial period, interests me and I want to comment on it here. I’ll be picking on Tableau, but it applies to quite a few software packages.

Thirty days seems to be the most common trial period for software. I think that is a reasonable number, depending on the type of software. Reasonable though it may be, it is the absolute minimum I will accept. Tableau and others offer 10, 14, or 15 day trials. Even fewer packages offer more than 30 days. Fogbugz, for example, offers 45 days. Certainly there is no one size fits all. It depends on the type of software.

One-off: This is software that most customers will use very infrequently. I burn probably one CD per year. If I needed CD burning software, I could probably use a trial and never buy the software. Other people may use CDs as an offsite backup and therefore use CD burning software every day. For me, the trial could be one day or one year and it wouldn’t matter. For others the 30 day minimum trial is appropriate.

Regular use: This is software that you use every day or on some periodic schedule. Tableau would fit into this category as would most Excel add-ins like Power Utility Pak. I won’t even discuss less than 30 day trials, so let’s compare 30 days with some larger number, like 90 days. I may be able to determine if I like an add-in in 30 days, but what is the cost to the software vendor for offering 90 days? If we assume that the sales are exactly the same, he will lose 60 days of interest on the price of the product. Can I go out on a limb and assume that a longer trial will mean more sales?

I think there are two factors at play here. First, the longer trial will reduce a barrier and cause more people to try it. “I have quarterly reports this week and I’m on vacation next week. Maybe I’ll hold off on downloading that trial.” With a 30 day trial, you run the risk of this potential customer forgetting about you, re-Googling, and finding your competitor. With a 90 day trial, this user feels like he has unlimited time to try it and won’t hesitate to download it. Second, potential customers who use your product longer will discover more features and more of them will convert to paying customers. Neither of these points is valid if your software sucks. If your product is bad, you need to pressure people into buying rashly. And the longer they use your product, the more bugs they’ll find. But I’m not talking to sucky software developers, only the good ones.

Lock-in: This is software that stores your data. It’s not necessarily that you can’t get your data out, it’s just that it’s more convenient to use the product that stored the data than to switch. Outlook, Quickbooks, and Fogbugz are examples of this type of software. Vendors should offer a minimum of 90 days on these types of products. In addition to more trials and higher conversion rates as mentioned above, the more data you store the more likely you’ll continue to use the product. I have 1.5GB of email in Outlook. If another email client is going to get me to change, it can’t just be better than Outlook. It has to be better to the degree that it overcomes either moving my existing email or living with having email existing in two different places. If I use Quickbooks for six months, how compelling would Peachtree have to be to get me to move all that data?

I’m sure there are situations where an extended trial period would costs sales. Development tools come to mind. If I have a six week bespoke software project and you sell a development tool that will help me with it, but that will not be very useful thereafter, a long trial would cost you a sale. Hopefully you know your market well enough to avoid those situations.

Maybe these trial misers are trying to create some feeling of scarcity around their product. If you offer a trial for you product, or you just have an opinion on the subject, leave a comment.

Posted in Uncategorized

18 thoughts on “Software Trial Periods

  1. You can buy a GM car today and return it now within 60 days. Nuff said. I still haven’t gotten to all the features of Supercalc.

  2. I personally think that a 30-days trial is more than adequate for a vast majority of software. I frankly don’t have time to be the QA tester for any software vendor, so if the thing doesn’t work in with 2 hours of me testing it, it’s toast.

    One thing I wish DK would mention is the partial trial. Where some trials have locked down features. Nothing turns cools my heals faster than trial software that has limited or locked down features. As soon as I see this, I’m gone.

    Camtasia and has a nice trial model. 30 days, full featured, create all the videos you want. Very smart.
    Xcelsius has also gone to a 30-day full feature model.

  3. On Tableau’s 14 day trial, I realise it can be extended if you ask, but I think Tableau are missing the point that you exceed someone’s expectations more if you offer something that they don’t ask for (but desire) than if you wait until they ask for it themselves. In economic terms, there’s an opportunity for a more pareto-efficient solution – a solution that makes both parties better off.

  4. I think it’s less how often you’d use a product than how familiar you are with the product category. Years ago when I had finally gotten tired of using DEBUG.COM as a hex editor it took me less than an hour with each candidate hex editor to like it or hate it. Same with utilities like Beyond Compare or 7zip.

    OTOH, you could give me a year with any photo editor or desktop publishing package and I still wouldn’t have a strong (or informed) opinion.

    That said, years ago (around 2000) IBM had a free trial for its PC version of APL/2 with a usage time limit, I think it was 100 hours or so. To me that’s a better approach than date-based periods (as long as you remember to exit when not actually using the evaluation software).

  5. I agree with all your points. The other kind of trial I see a lot is one where you get unlimited time, but key functionality is crippled. I don’t think these are very effective, because if you’re not using the app. properly, you’re not really trialling it. If you can’t save your work, say, then all you will day is tinker with an application.

    Re: Tableau’s 14-days.
    When I was trying to get my company to purchase Tableau, we extended our trial over and over again. Probably to at least 90 days, maybe 120 or more (I work at a 700+ yr. old university – things move slowly!). Tableau extended the license without grumbling each time. I personally never found it a problem to email them to get an extension; one email was all it took each time. I accept that some people may.


  6. I thought about mentioning crippled trials, but the post was already too long and boring. I didn’t think about timed trials or per use trials, but both are worthy of discussion. I guess I should do a part II.

  7. I offer a demo version of my Reftreeanalyser (, which has both limited functionality and shows a small nag screen when opened.

    Opinions please: what would you prefer: time-limited with full functionality, or limited functionality with nag screen or other??

  8. I prefer long trials because, frankly, I’m lazy and don’t always get around to properly evaluation the app. When I do get around to the app and it has expired, I’m more likely to just forget about it. I think full trials which convert to limited functionality is probably the best. If I can still use some of the app I’ll tinker with it and give it more of a chance to win me over.

    I don’t see the advantage of a short trial except to maybe force a decision. I would think hobbled app’s kept on more pc’s will eventually lead to incremental sales and at zero cost are effective. For example, I just bought a license for NoteTab after keeping it in its hobbled state on my pc for a year. I used it rarely and when I needed more, their it was — reliable with a reservoir of goodwill.

  9. I have to make a comment seeing as it was I that originally stated that I thought Tableau’s 14 day evaluation period was too short.

    I still do think it is too short. This limit was enough to stop me doing an evaluation when I first encountered Tableau (Andy, was it you telling me at the Excel gig?). It is an expensive prodct for me, no matter what anyone says about RoI I won’t get it directly. It would be something that I would look as part of a solutions package for clients should they need that richness, and also be up for that cost. So, if I am to evaluate it, it is not just like a hex editor, fire it up and see if it works in a manner that I like. I have to be able to play with it, understand how it works, where it works best, what its limitations are, and so on. Trying to earn a living does not make that easy with a complex product (I hope its complex for that price) like Tableau, and I don’t think I can do it in 14 days. As Jeff says, extending it is fine, but turn us on not off, offer more in the first place.

    @JKP – not crippled functionality please, as Mike says I don’t even bother with these.

    @Andy – you stated that kept extending … ‘when I was trying to get my company to purchase Tableau’ … What made you want to get them to buy it (in other words, without having trialled did, what made you think Tableau was for you)? And what was your company’s criteria for saying yes? Either way, you are arguing for longer trial periods.

  10. So far my shareware add-ins have had only a 30-day trial and I have not had a request for an extension — but then evaluating an add-in is pretty straightforward.

    Recent developments have caused me to reevaluate this “unlimited features for N days” approach.

    One popular add-in is TM Timer for PowerPoint – One of several different uses is at conferences where the presentations are meant to be time-limited. Someone contacted me a few weeks ago with a series of questions and concerns, all of which I politely answered. I suspected — and sadly, my suspicion was correct — I would not get a sale because she indicated the add-in would be used at a conference “next week.” The last I heard from her was the last of her emails asking for some more clarifications. Since the 30 day trial included the conference apparently she saw no reason to pay for the add-in.

    Another popular add-in is TM Randomize Slideshow – There are customers who use the program on an ongoing basis (Flash Cards for rehab, randomized slides for class quizzes, etc.) but there are also those who want it for a “one off” use like a memorial or a party slideshow. Those in the latter category — including this one person who worked at a museum and at whose request I even added new capability — typically fall in the 30 day trial period and therefore apparently don’t see the need to pay for the add-in.

    Someone at the IT department of a law firm contacted me about TM Directory ( asking about a site license because they wanted to use the add-in on almost all of the computers in the company. Over a few emails we worked out the details and I also clarified questions he had. He indicated the add-in would be deployed starting “tomorrow.” That was the last I heard of it. Either the company found another solution or decided not to pay for the add-in since the 30 day trial covered whatever it wanted to do with the add-in.

    I am actively considering modifying the definition of a trial. One option is to use the minimum of N days and/or M uses and/or P hours of use. But none of them will address the “one off” use.

    One solution is to add a watermark or a splash screen as might be appropriate.

    Another — and if I do anything I am leaning towards this — would be to limit the scope of the add-in. All features would be available simply not on an unlimited basis. For example, the TM Directory add-in would be capped at X files, the TM Randomizer would randomize a maximum of Y slides, and the TM Timer would be active only for the first Z minutes of a slideshow.

    I may just leave things the way they are but it is frustrating to know people in a range of organizations are so unscrupulous that they will rip off the work of others on behalf of their organization!

  11. Interestingly enough I have Both Tableau and Xcelsius. I will keep my comment to Tableau. Did I like the trial period. Yes and No.

    Tableau is real quick off the line to get using and if you are used to XL then this software is not a stretch. Where the trial period falls short is that it assumes you have all of your source data in good form. I tried hooking directly back to our DB2 but that proved problematic as the query times were way too long. To be efficinet I had to get an Analysis Services cube set up. That took a little while to get done and by the time it was ready the trial period was darn near over. As luck would have it I did see enough potential that on my recommendation the company sprang for a licence but it was tight. I was in the middle of a project where we could get very high pay back if I managed to find a coulple of things (which I did).

    The value of tableau is in wading through huge volumes of data. A longer trial period to get your data in order would be appreciated.

    What is the right trial period. How much time will it take you to sufficiently kick the tires. If you download it and don’t try it for six months then how interested were you in it in the first place?

  12. Hi Jim. I absolutely agree that alonger trial period to get your data in order would be appreciated. First thing I did is cut some really complicated SQL queries out of Excel that feed a Pivot Table, and paste them into Tableau. They didn’t work, and I haven’t had the time to look at them. Maybe this will only take 5 minutes to fix, maybe it will take 5 days to track down the issue…who knows. What I do know is that I’m struggling to find even short windows of time to evaluate it at present, let alone hook it to datasources.

    Mind you, Andy Cotgreave makes a good point over at the Contextures post that data soucre issues can be ironed out with time, and that in the meantime by connecting directly to data in Excel a user can make good use of the trial, by seeing how simple it is to create vizualisations, and the rapid way you can find the story in your data.

    Or as he puts it: “Ironing out connections can come later; in the meantime, just hook Tableau to some of your Excel files – I am sure you’ll see the power of Tableau that way.”

    A good point.

  13. Dick there is an argument that a longer trial gives a lower conversion rate as people lose interest and never buy – the long tail – but an empty one.

    The trouble is people are now confused between *trial* (to try the product to decide if they want to buy) and *free use period* (a 30/60/90 day period where you can use the sw to do whatever before deciding if you should pay to *continue*). And lots of people don’t buy during the trial – they wait till the end, or don’t buy at all.

    For many utility type stuff 20 mins is probably enough to see if its worth 20/40/60 dollars. Seriously who needs 90 days worth of cost benefit analysis to authorise the capital procurement of a 20 usd/gbp tool?

    I totally accept for larger more powerful products different issue apply.

    My trials have reduced functionality and the odd nag. These tools therefore always have some value. A time limited trial thats out of time is worthless. And I have ended up with loads of these as I half review something then get pulled onto something else. If its expired by the time I get back to it I just uninstall and move on. In fact I rarely even download them now I’ve been burnt so often.

    The instant worksheet password remover (trial) is one of my most popular downloads. The trial has a few nags and can’t be automated – but it shows how fast it removes passwords – the key selling point. Sales are rubbish, but at least every time someone uses the trial they get a little advert for Codematic. (Sales are about 0.5% of trials.)

  14. Maybe I’m dumb, but I don’t have a formal trial version of any of my utilities. I offer downloads of the older free versions of a couple, so people can see how they feel, if not try out all the features. I have thought about including 30-day kill routines, but that seems like a lot of work.

    I have 90-day guarantee, and I don’t give anyone any crap if they want a refund (even in one case beyond 90 days). I haven’t had to refund many purchases, and it’s not worth the hassle to fight it.

  15. I often downloaded stuff, installed, got distracted with real work and then was locked out.
    Now I might download it, or people send it to me for review, but I don’t install until I have time; so I tend to just forget it.

    Time limited trails are hackable in addins, but not with compiled code.

    How about offering different versions:

    1) Free – does one or two things, shows the other options it can do but just says ‘available in paid versions’

    2) Personal – does what an end user would want to do mostly

    3) Professional or Premium – with all the geeky stuff that looks shiny but is only used occasionally.

    30 day warranty sounds good, 90 is better, I don’t expect more.


  16. My favourite was a fully functional version of something (can’t remember what now) with a nag screen that took longer and longer to clear each time you ran the software. It worked: I bought the licence because I’d started to use the product. But that was for a usefull tool.

    (Almost) fully functional is critical – I’ve given up seriously thinking about several applications that were crippled long before I’ve decided if they are useful.

    I would not worry unduly about “one off users”. So what? They are not going to buy anyway for that one use. At least if they demo it at a conference (referring to post above) somebody in the audience might …

  17. I don’t yet independently publish software (working on it–thanks PED II, and all the rest of the fabulous information out there) When developing an Excel tool for myself or another department (I’m a health actuary), it needs to be fully functional. I think time-limited-with-full-functionality trials are very much better than partial functionality trials. I’m too curious and I want to kick all the tires and drive the car on the highway before I buy.


Posting code? Use <pre> tags for VBA and <code> tags for inline.

Leave a Reply

Your email address will not be published.