How long should your “trial” period run?

I’ve had this running debate with a handful of friends – I’d love to throw it out there for comment. The questions at hand are 1) whether companies should offer a “free trial” period for their software/web service; 2) if they do, how long should it last; and 3) what information should you ask for before starting a trial (specifically should you ask for credit card information up front).

Here are a few thoughts. I’d love to hear your opinion.

While you know that your web service is the greatest thing since sliced bread, it’s really really hard to convey the chocolaty goodness that is your product to the average consumer. You have to pull them in and offering a dry run of what you do is a good way to do this (maybe you have a free version of the service as well, but even so, you’re saving the best features for your paid users and you need to show them what can do).

Now on to the question of how long your trial should last. While avoiding the obvious cliche answer (“as short as possible”) I’d point out that many (most) companies default to 30 days. I think this is a mistake. If you can’t show value in a week or two you’re doing something wrong. And by waiting 30 days you’re just extending the number of people in your trial funnel and making more work for everyone involved (not to mention stretching out the trial and potentially losing customers). Your product should be designed to quickly get people up and running and to show value right away (see some ideas on that here). I believe your trial period should have three parts: 1) onboarding – quick and as painless as possible; 2) show value – make sure you’ve designed what you do in a way that you’re in front of your customers immediately after they input information; 3) ask for the conversion to paid – once you’ve shown them what great value you add, ask them to fork over the dough. Most companies err in all of these categories. Onboarding is too difficult. Value comes over time, not right away and the product forces users to remember to come back to it rather than the other way around. The “ask” comes too late, after people have forgotten what the service did in the first place.

Lastly open for debate is the question of how much information you should collect up front. On the one hand are people who believe that you should collect as little as possible – probably just and email – and get people into the system and up and running. On the other are people who believe that since you have their attention you should grab their credit card information up front so you can start charging if they don’t opt out. I’m in the former camp. I couldn’t find any scientific study on this, but in my experience asking for a credit card (or even having a two page sign-up form) significantly drops the number of people who get from the beginning to the end of the sign-up process. In one case, asking for credit card information resulted in 9 out of 10 people who started the free trial process dropping out.  That’s not worth the back-end trade-off of having the information to charge them later (plus the opt out thing isn’t cool in my opinion).

Your thoughts are welcome/encouraged!

  • Seth,
    Good to see you blogging again! An interesting topic, and one that many organizations like ours continue to debate. We have a SaaS solution that requires about 3 minutes to implement for prospects and customers. We have had 30 day trials, 21 day trials, and now are at 14 days. And I don't see us going anywhere but down on that amount of time. Two weeks is plenty of time to work with prospects who already have an interest on some level of your space (in our case, marketing automation). With the ease of seeing our product in action, the next step of showing them value as it relates to their business needs is the focus for the next week or so. After that, you ask for the business. Its that simple.
    I'm sure if the solution was more complex in setting up there may need to have longer trials, but with most SaaS solutions you should not have to have extended trial periods in my opinion.
    Thanks for the post, Seth.

    • sethlevine

      Thanks Matt. I got off the blogging wagon for a bit and really missed it (too much travel!).

      I appreciate your comments. Do you have any data you can share about conversation rates (or even relatively conversion rates) as you pulled in your trial period? I think your experience is pretty common – you start with some period for your trials and then realize you can pull them in…

      • My pleasure, Seth. Our conversion rates (to paid customer) were at or near 50% when we made the adjustment(s). We felt this was far too low for a product such as ours that easily shows value in short order and is at a manageable price point that doesn't requires layers of authority to sign off on. Now at 14 days, our conversion rate is over 70%. Who knows, we may go to 7 days!

        • sethlevine

          wow – that’s great (and compelling) data.

  • A limited-time trial is one of the major freemium models, right? (Charles Hudson has a pretty classic post on time-limited vs. feature-limited freemium, btw:

    It's a great question. I've heard of better results coming from both a switch to “collect cc later” and “collect cc earlier.” With inconsistent results, I'd say you have have have to test it…which, these days, is about as unhelpful as advice can be. I'd love to see some results on significant value-data sets.

    • sethlevine

      thanks for the pointer to the hudson article – interesting insights. clearly this is a debate that has many dimensions. . .

  • HayNow

    Replying as a consumer of these services and having bought several (or converted to paid versions):

    30 days is usually good, 15 minimum.
    No credit card right away – lower the bar.
    Possibility for month to month (go back and forth between paid and limited free)
    Lower the bar.
    Lower the bar.
    Have great, reactive support early on, before they're paying.

    All the theory in the world can't compete with the fact that I am paying some people for services and others didn't get me because the trial was too short or they needed a card too soon. In one case I paid, asked for a fix (something wasn't working) and got nothing. Had to ask for a refund.

    One other important factor: reach out to the trial consumer, ask them if they need help to see your awesomeness or if they have suggestions during the trial.

    Make it dead simple to send feedback. Eschew Get Satisfaction that puts too many hoops in the way. Just get the comments and filter them.

    Or maybe I am atypical, “Let the market decide”.

    • sethlevine

      your experience sounds pretty typical and this perspective is helpful. a good reminder that one way to make these decisions is to talk to a group of existing or potential customers (which companies should be doing in any event).  one thing i didn’t point out in my post but which your comment made me think of is that one other benefit of a shorter trial period is that those that have a good experience become customers faster and for those that are having difficulty you can always extend the trial period. if you’re trial period is 30 days (or longer) you’re not going to find many people buying early and you may completely lose the 2nd group before you even realize there’s an issue.

  • AlisonW

    The basic issue (as I see it) is that offering a free trial shouldn't create the position that the prospect doesn't then need the 'real thing'. If I am seeking a solution to something and I find a 'free' trial lets me carry out the one-off task adequately then I'm going to wonder whether I ever need to have the paid version.

    The reverse, of course, is that if the limitations on the free version are too strong I might not get enough from the trial to make me believe that I require the paid offering.

    • sethlevine

      i agree. and would add that not only should the trial not undercut the need to buy the product, it should reinforce it. i.e., the experience needs to demonstrate almost immediate value so you’re really drawing people into the product. a week later you want them thinking “how could i live w/o this…”. the Pandora integration to sonos comes to mind for me personally (of course in that example i’d probably pay them several times the current annual fee for the service)…

  • All depends on the product.

    Take the burgeoning WPThemes business. I'd suggest a free model but focus on a subscription or one-off payment to access the support/community. Regular theme updates are a must. Great product development and encourage community involvement with showcase sites, etc. But give it all away. People will nab the stuff off torrent sites anyway.

    What will be priceless: a strong community forum, answered email support, product innovation … and lastly, listen to your customer's needs and wants.

    • sethlevine

      there’s a long post to be written at some point on what fee models work best for what kinds of businesses (freemium, pay for every level of product, free product but charge for upgrades, free product but charge for other things in your service, virtual goods, etc.). i’ll give that some thought …

  • Besides credit card, what other ways have you seen for a company to try and limit trial periods/accts to “one per customer?” In the gaming industry – esp for multiplayer games – trial accts usually come w/ some base grant of virtual currency or other in-game value. Companies which rely upon solely an email address often find themselves victims of trial acct farming for virtual items which can be sold for real dollars.

    • sethlevine

      typically email (people can obviously have multiple emails but at least it’s a little bit of a barrier). the real key is that your system should be set up so that gaming it with multiple accounts really doesn’t get you all that much. in the case of multiplayer games, because you’re earning points and experience there’s limited reason to set up multiple accounts. and in any event i strongly believe that industry is moving towards free game play with intragame opportunities for the game maker to make money by selling virtual goods, etc.

      • i also found a few folks using mobile # w/ SMS confirmation as a few ways to ID uniqueness during sign-up

  • Definitely ask for as little as possible up front. Foot in the door technique.

  • Jeff Yoak

    Perhaps it is lame to say “It depends” but I think it does. I'm using a demo of a Geometer's Sketchpad right now that is 90 days long. I originally downloaded it to visualize a particular demonstration implemented with it. On a short demo, it would have expired and I would have forgotten about it. With a longer demo, I find myself returning to it from time to time. I'm starting to think that I may want to keep it around. There have been other pieces of software that I wasn't sure would fill my need, but I needed about 5 minutes of demo to see how it worked, and then buying it was certain.

    It seems that the demo period ought to be long enough to allow you to see the real value, but not any longer than that, in order to create urgency. I realize that this will vary from user to user and you have to guess in any event, but it is important to remember that it may also vary widely depending on what the software or service does.

  • Very smart post. Thanks for the insight.

  • This is a great topic! I'll add what we've been doing at

    At my company we've tried it all! Our original trial period was 30 days with only a name and email requirement and when a customer would ask we would extend it indefinitely. This allowed us to really seed the market with trials (good) but made it hard to sell the product and follow-up with trial users (bad).

    And so the experimenting began:
    1. We dropped the trial down to 15 days and forced the user to offer up a little more information that allowed us to follow up directly once they downloaded. At the same time we added email verification — sending the trial codes to a valid email address.
    2. After a while, we dropped the trial down to 7 days at the behest of sales (“we’ll close the deal faster”) which was a disaster! Downloads didn’t suffer but we certainly received our fair share of people calling for extensions.

    We finally settled back on 15 days with extension requests coming directly to the sales team who would extend for days, not weeks. They key for our product (EVERY product for that matter) was that we needed to make sure the end user actually had a need to USE the product during the trial and found value in it. This is a very important and often forgotten piece to this puzzle.

    Whatever you do, a process for follow-up is as important as the trial itself (as has been noted in some comments already).

    Hope this helps!


    • sethlevine

      great data rob – thanks for adding to the conversation!

  • Presci3nce

    Great question re trial length. I own a software company and we provide 7-day… I found your page because I was trying to find research regarding optimal length, i was thinking of reducing to 3-day.

    • Frank

      I also own a software company, and we too considered a shorter trial period, but most companies balked at the idea of a 3-day trial. We have used a 7-day trail period for nearly 5 years and it has worked out well for us. We only deviate on a very strict exception basis (pronounced “if the deal is big enough”).

  • Frank

    Seth, I could not agree with you more. My company offers a 1 week trial period, and as part of the trial we need to do a little work from an onboarding perspective (configuration, training, minimal data, etc) in order to create a truly valuable user experience, but it has proven to be worth the effort, as most users are hooked once they have an opportunity to see and feel a configured solution. The reality of a B2B solution like ours is that a user already has a full-time job, and they don’t have the time or inclination to focus on a trial for an extended period of time. We have been tracking the trial period usage for a little over 7 years now, and it drops off dramatically after the first week anyway. People want to see that the product does exactly what they need it to do. That can usually be done in a couple of days, so a week has proven to be plenty for our trials. As far as information that is captured, our business is a bit different, so credit card information doesn’t really apply.

    • great data frank. so for you guys there’s a serious drop off after 7 days? i have to say that’s my preferred trial period for most products…

  • insert1236

    It is an informative post. Who believe in other people, because you have their attention, you must enter your credit card before you can start charging if you do not quit.

  • Anonymous

    I think this is a mistake. If you do not see the value of a week or two you are doing is wrong. Get people to price their products quickly and should be designed to appear immediately. nnCustom Table Pads

  • Anonymous

    Then there are those who want to because you have their attention. You must enter your credit card information in advance. So you can start charging if you do not give up.

    good gifts

  • 1) I think free trials are a win – win. Bottom line – I don’t think anyone wants to be in the business of winning over customers that don’t really want and value their service. If you don’t offer a free trial, you just end up with unhappy customers who could have figured out it wasn’t a good fit sooner.

    As a consumer, this is one of my biggest frustrations with the iOS App Store – so much so – it gets in the way of me buying apps I probably would otherwise.

    2) People need deadlines. If you solve part of their problem for Free – for ever – too many just won’t pay to have you totally solve it completely.

    3) I can’t see the opt out trick really being sustainable anymore. A scalable business is built on the foundation of recurring revenue. It’s not the first transaction that matters, it’s the one after that, and after that, etc. Relying on people forgetting to cancel is not serving anyone.

    First impressions are just as important with a new customer as they are in all other parts of our lives. Why risk tainting it with a trick?

    • Your app store comments is an interesting one Nate. I find myself very reluctant to buy the more expensive apps for exactly that reason – there’s only so much I’m willing to “risk” on a crappy app that I don’t end up using.

      Good comment all around. I wrote this post quite a while ago, but with more services charging on a subscription basis I think the ideas here are as relevant as ever.

  • Jesper Faurby

    Very interesting debate.

    I think it depends on what kind of service you offer. We run a SaaS b2b service and we are considering going in the other direction and offer a 60 days free trial period!

    It will give our potential clients more time to test our software and hopefully get them “hooked”, so that they are happy to pay after the 60 days. Instead of chasing them in the usual 14-days free trial period.

    Also think it might be a competitive advantage to offer a longer trial period than your competitors.

    Of cause from our point of view a short trial period and credit card information in advance would be perfect … but I don’t think it´s the best in the long run. Not for our trial users and not for us.

    Any comments?

    • The challenge with really long trial periods is 1) you run the risk of asking for the buy past the optimal engagement period (and people tend to engage more if they have put down money for something) and 2) often what your’e really doing is just delaying revenue from people who are already ready to pay (but without increasing your conversion percentage). All things that can and should be tested. But my experience is that anything over 30 days is just leaving money on the table – if you can’t convince someone in a month, you’ve lost them. Better to put them in a re-engagement funnel than pretend they’re still in a trial funnel.


      • Jesper Faurby

        Hi Seth

        Many thanks. I really appreciate your input!

        But I still think it depends on what kind of service you are offering. We are BtB and are usual going in to long time relationship with our clients after a successful trial period.

        We are often experiencing that our trial users are too busy to test our software within the 14 days, and are asking for more time.

        Also you can limit the Trial users access to certain features, and in this way give them an incentive to sign up for a paid version, before the trial period expires. A kind of a 60 days freemium plan.

        E.g. Basecamp ( is now offering a 60 days trial. And I’m quite sure that they have run some tests before extending their trial period from 30 to 60 days.

        If our competitors are offering 7 days trial, and require credit card informations in the trial period. And we are offering 60 days trial and no credit card, I think (hope) that would give us a competitive advantage when users needs to decide which solution to try out.

        Anyway, I think we’ll try it out, and I’ll return in this thread with our findings.

        Thanks again.

        • I’d love to see the data/hear how it goes for you. Can you drop me a line (or post here)?

          • Jesper Faurby

            Hi Seth
            Sure, I’ll let you know our findings 🙂

          • Hi Seth

            Hope you are well.

            I promised to return in this thread with our findings when changing the trial period from 14 days to 30 days.

            We have now been running with 30-days trial period for about 10 months. It a bit difficult to see exact data about the impact for this specific change, because we have changed a lot of other things as well.

            But we think that it was the right decision for us for the following reasons:

            • More people signs up for a free trial now
            • Our conversion rate from trial > paid is the same as before
            • Not so many ask us to extend their trial period because of vacation etc.

            I think it’s very important to stay in contact with your trial users, especially when extending the trial period. We send them some personalized (but auto generated) emails with tips and tricks etc.

            All the best,

          • Harald Stelzer

            Hi Jesper,

            I followed the discussion and hadexactly the same oppinion as you. But as I see, you have changed the trail period back to 30 days. Would you mind to tell me your reasons?

            Best regards