Loquax Research on Twitter Competitions

Posted on: February 10th, 2011 by Jason 35 Comments

Twitter competitions have become increasingly popular over the last year or so. Every day, especially Friday, there’s usually a number of giveaways offering prizes ranging in value. To enter an entrant is usually asked to retweet a message to enter. What we, at Loquax, wanted to do was take a look at Twitter comps and see if we could find out more about them.

What we wanted to see was how many people entered, do many people enter more than once, can people who enter more than once win, and are winners randomly selected.

This is the very early stages of our research, but we thought it’s a good time to let you know what we get up to beyond what you see on the competition listings.

Background

To do this we selected at random a number of competitions from various promoters. They range from big brand fashion sites to gift shops. For now we’re opting to keep their names hidden as we think that’s the fairest thing to do. Our criteria for analysis was that the competition needed to have a named end point (e.g Comp Closes at 5pm) and the competition had to be a “retweet to enter” mechanic.

The first issue we encountered with our analysis was dealing with Twitter’s search set up. Twitter’s system can be a bit of a nightmare to search and locate all entries – especially if the competition lasts over a long period. In fact, we’re quite concerned about any retweet competitions that last longer than a couple of days. Unless promoters are databasing entries as they go along, we’re unsure how they pick winners given the quirky nature of Twitter’s API (advanced method for searching essentially).

Because of this issue our analysis has been restricted in terms of historical competitions. By historical we mean as early as last week. Therefore data really can only be analysed from competitions that have run over the last few days. It may well be the Twitter search is playing up, but our feeling is we have these issues then maybe so do promoters?

Number of RT to Enter Entries

Once we were able to do searches using Twitter’s API system we started to find some intriguing data. The biggest surprise was that the number of entries in a Twitter competition seemed to be very low. Remember that these comps tend to only be for a few hours, but even so we were surprised to find many attracted between 200 to 500 (maximum) entries.

We also analysed when people tweeted – breaking down the time of entry compared to the original competition tweet. Most people tended to enter at the very start of the promotion – with the first hour or two tending to attract the bulk of the entries. However is this a good strategy?

Best Time to Enter

The good thing about Twitter is that most promoters name their winners and so we were able to look at who won and when they tweeted. Our initial assessment is that promoters tend not to use automated randomised methods to locate winners. Only one company, based on the format of their winner announcement and the time (early hours of the morning), seems to be using an automated system.

This means that winners could well be picked by a “pin in the search results” method. Not the most ideal. Of the competitions we analysed a lot of winners came towards the end of the competition time. In fact only one competition of the twelve we looked at had a winner who had tweeted in the first hour of the promotion. The majority of winners always came in the second half of the promotion time.

Can Doing More Than One RT Win?

Another interesting fact that cropped up was the one of duplicate entries. Every competition had duplicate entries! The vast majority of entrants only entered once, but roughly 10% make 2 or more tweets. And entering more than once doesn’t seem to stop people winning.

Of the competitions we analysed half of them had winners who had retweeted more than once. This clearly indicates that promoters either don’t have rules about number of entries that can be made OR they’re not checking to ensure all entrants are playing on the same level field.

Early Conclusions

Of course all this is preliminary research but it’s quite eye opening in many respects.

The number of entries in a typical RT to enter competition varies but it’s not as high as we expected. People who enter more than once don’t seem to be penalised as they can still win but the vast majority of people just make one entry. Entering at the start of a promotion is popular, but we think there may be more success if you enter towards the end.

Of the promoters we looked at only one seemed to be using an automated method of picking a winner (based on their time/format of announcement). However, our early understandings of the Twitter API system seem to indicate that managing a long term twitter competition could be quite difficult and compers should be wary about how winners are picked under those circumstances.

We’ll be looking at more Twitter competitions over the coming weeks to see if our early analysis has foundation and also bringing together some data and pretty graphs to support are findings.

  • hamishmash

    I run these type of competitions for a small cosmetics brand, beautyuk (@beautyuktweets) and have been getting some good results really considering the small amount of effort you have to put in to create them. And because our prize is directly related to what we sell we are also finding that about 85% of entrants are our target market. We have just set off a new competition which is the first one we have done in conjunction with another brand so we’re interested to see how that goes! To monitor all our entries we use the social media monitoring tool Brandwatch as well as receiving Twitter mention alerts from Twilert. This makes sure we can have a full history of everyone who enters. We then use a random number generator to pick the winner.

  • JSH

    That’s really interesting I will be keeping my eyes peeled for your findings. I have entered lots of Twitter Comps lately but had no luck and beginning to think I could be using my comping time more wisely. Saying that, I have seen some big wins this week and it only takes one tweet! I usually enter straight away but not sure how I would remember to do it otherwise, also a lot of them don’t specify an end time.

  • cpjw

    Excellent article, thanks.

  • fitwell

    very interesting have been doing twitter for a while and only one win…thought I must be doing something wrong

  • timmmera

    So, all day the numbers entering are lower because the majority are working and too busy or unable to enter comps…most comps end around tea time or early evening to allow those people to get involved…which means that many more enter in a short time later on with the obvious result that wins are more common for that larger group.
    Also, the people around during the day are short of people to inform them of the comps compared to the later entrants who have a whole day of retweets to go for additional to their own group retweeting.
    How could it go any other way with this type of comp but the larger group wins more and as that group is the ones coming home from work that means later entrants win more. Nothing to do with the draw, just basic pobablility.
    Youll prove this with week long comps, either on FB or twitter.
    This article is leading people to think there’s some mystic reason that they don’t win entering early, that’s twaddle. The sample size is laughable too, based on a sample of that size I would be eating burgers 50% of the time fot lunch because I have done 3 days this week when in reality I rarely eat them.
    Far from being excellent it’s very disappointing, you may as well have blamed it on starsigns that people win more or less lol

  • timmmera

    Another thought.
    Imagine a day long comp, all day people enter and RT (some more than once since that seems not to matter) and that accounts for 200 entries, then the mob come home, with no time left to multi RT they make up 300 entries due to sheer numbers.
    So we could have 100 people in the early group vs 300 in the later one…and no way the later group wins propertionally more based on that set of stats. If it’s even close to even, say 60/40% and not 2:1 the earlier group have a massive advantage over the late comers even if the latter group wins more comps they aren’t doing as well as the first group who should only win 1/4 of the time based on 400 total RTs. and a ratio of 100/300 between the two that produces. 😉

  • @timmmera – good to spark some discussion, but you miss some relevant points. We quite aware that it needs further analysis as we do say quite early that “this is the very early stages of our research” – “of course all this is preliminary research but it’s quite eye opening in many respects” – “we’ll be looking at more Twitter competitions over the coming weeks to see if our early analysis has foundation”.

    We’d like to do a larger sample size, but having played with Twitter’s API believe me it’s not the easiest. We struggled to get some data out (as mentioned)- so how do promoters manage it for longer time period competitions?

    Your probablity comments are logical for an average free prize draw and if the pattern you indicate is correct.

    However as mentioned most people enter at the start of a Twitter competition (”Most people tended to enter at the very start of the promotion – with the first hour or two tending to attract the bulk of the entries”) – yet of the comps we looked at it showed that people win from when they enter later – when less people are entering (which is opposite to your thoughts above)

    It’s all well and good jumping on us and saying it’s disappointing, but you’re being quick to dismiss something without reading the various points made (imo).

  • linheiner

    Thank you for doing this research – I’m really pleased that you’re doing this and am interested in seeing more. I have been doing twitter comps for a while now with a few wins. What constantly amazes me about their comps is that even though the entry numbers are small the same names crop up again and again as winners. One person in particular stands out as winning very frequently, often 3 times in a day, which seems beyond the realms of statistical probability. It would be very interesting to see if you come up with a “winners’ profile”!

  • And I should add that the majority of the competitions we looked at finished around 4pm or 5pm – and tended to have a <8hr shelf life.

    So our initial findings are fine - but as mentioned need more study. But if you have any data contrary we're more than happy to discuss it - as we've said it's a first look into these things, the issues we had at running historical data and what we're initially seeing.

  • Ennill1

    I did mention to one site on twitter that the same names seem to come up quite often as winners and they replied:

    “Most comps on twitter want you to enter as many times as you like. Good to spread the word 4 company x ”

    They said they used randomizer to choose winners, but I wonder that if you RT x no of times if that qualified for one entry, or if you RT 50 times, that would be 50 entries? I also mentioned that I thought that twitter comps gave limited publicity as many people can’t spend all day on twitter and the reply was

    “Don’t really know how it works. I guess that we have so many followers who have x followers but not all on at same time? ”

    I wasn’t having a winge at the company, just being curious.

  • @linnheiner – “It would be very interesting to see if you come up with a “winners’ profile”. One problem we have is that a lot of twitter comps randomly start and randomly end… so it’s hard to work with the data (at least at the moment). Also twitter comps that require an answer or funny reply are hard to track. We will though look further into things.

    @Ennil1 – “Don’t really know how it works” – a good point and if they don’t know how Twitter works, how do they know how to fairly pick a winner from a search system that isn’t too good for historical data?

  • Loupri

    Hmmm very interesting views, I look forward to your further analysis Jason.

    I’ve been entering Twitter comps since August and I’ve won a grand total of one prize 🙁

    Wow, maybe I should spend my time entering other comps and forget about Twitter!

  • Ennill1

    One UK site says on twitter:

    “our competition is now closed. We will now select a winner using http://twitrand.com/ ………… ”

    They chose a winner – but twitrand no longer works…. so I wonder what method they actually used…

  • @loupri – “maybe I should spend my time entering other comps and forget about Twitter!”. As pointed out the dataset is small and to that extent limited which means it’s not enough to say don’t enter ever. Remember these are just RT to enter comps and there are other twitter comp types.

    What I would say is have a go at trying to RT towards the end of the promo (e.g. if it ends at 5pm, try entering at 4pm rather than at 10am when the first message appears).

    It may well that next week our data will show something more random (as you’d expect).

  • libra100

    Very interesting discussion.

    I get very little time to spend doing Twitter and Facebook competitions. Interesting that the initial stats show that winning seems to be skewed towards later entrants. The only problem seems to be that it’s difficult to ‘save’ Twitter competitions to enter later in the day, as you can with the usual website competitions.

    Thank you for the research.

  • gaynorj

    I enter very few Twitter competitions, but your findings on the low number of entrants might encourage me to enter more….. even if it does do my head in 🙂

  • lebeeuk

    In response to JSH’s comment about how do I remember, what I’ve tried a few times in the last couple of days is to favourite the post & then retweet it a bit later in the day.

    I’ve been off work sick for the last week so have spent a lot more time than normal on there & like some one else has commented there do appear to be a handful of people that keep turning up as winners & even when the T&C’s state to only tweet once or once per hour they appear to be taking no notice of this. This was queried by a number of us regarding the winner of a large prize the other day yet the company holding the competition has completely failed to come back with an adequate response even though it was their own T&C’s which had been broken – I think this comes back to random tweet selection & no notice being taken of the physical data.

    Companies that suggest the multiple retweet is a good option for them are also failing to realise that the people who spam tweet will actually lose followers rather than gain them & therefore the publicity would be reduced?

    It will be interesting to see your findings Jason, especially after yesterday where many people were losing their twitter feed in the middle of the afternoon.

  • clareandtribe

    my partner and i have decided to try twitter me only starting today and my partner two weeks ago it seems alot quicker way of entering comps as my partner can enter roughly 80 comps in an hour using the program tweetdeck which would take me nearly 4 or 5 hours to enter the same on loquax, ive also signed up to a app for my android phone called tweetcaster which roughly works the same as tweetdeck and i can enter lots of twitter comps with ease using it, i see you can add two or three acounts to the tweet deck program and this allows you to only have to retweet once but it counts on both mine and my partners account so in essence your entering twice but only retweeting once (if that makes sense) i havent had any wins so far only having used it for less than 24 hours but my girlfriend has won 4 prizes in as many days only small wons but a wins a win,

    very interesting reading jason thank you

  • steeps

    That’s really interesting. I’ve tried to use the Twitter search facility to get a rough idea of how many people are entering comps and whether or not people are making multiple entries but it’s nigh on impossible. I find it hard to believe that promoters are keeping a database of every entry as it is made and therefore it’s much more likely they’re just picking winners from the most recently visible entries.

    Having said that, I’ve won a handful of twitter comps (I don’t have time to enter many), and I tend to RT when I see the comp appear, and only occaisionally do I make multiple RTs. (If a competition runs for several days I would probably RT once or twice a day unless there are T&C’s that state otherwise – and there rarely are.)

    I’d like to see more comping statistics – not just for Twitter. It’s interesting stuff!

  • suenatal

    interesting reading Jason, thanks for taking the time to look into it. I’ve only been tweeting for just over a week, but have had a win. I find the most difficult thing is remembering whether you have tweeted it already, especially if it is a one entry tweet comp. Getting that many retweets its too difficult to follow. At the weekends in particular I can log on after a couple of hours to have 600+ new tweets which are impossible to read.

    Look forward to reading more of your findings

  • Alexbe34

    I have used RT apps as don’t have a PC at work or a smart phone so miss the short expiring competitions .Also had some success with a few too .
    From what I can gather from this site & some other comping sites that I visit more companies are doing Face Book or Tweeter competitions as they are so simple too enter.However the drawback is some tweets running are USA only or will require an import tax payment if your lucky enough to win 🙁

  • Loupri

    Actually, thinking back to the one win I have had on Twitter, I entered in the last five minutes of the comp ending! So I’ll definitely try entering closer to the end of them from now on and see what happens.

  • Jo Haigh

    An interesting investigation…I’ll look forward to further blogs on this. 🙂

  • NanaVal

    Gosh that all sounds very interesting and I look forward to hearing more on the subject. I’ve not signed up to twitter as yet as I still view it suspiciously (as I did facebook when it started 🙂 ) but I might have a look at it. I do enter a few facebook comps now, but not very often (no wins so far) as I tend to concentrate on online comps mainly. Perhaps if it’s a comp you can enter more than once you could enter at the beginning of the comp and again towards the end to up the odds.
    Thanks Jason – very interesting article.

  • KIERAN WALSH 1949

    very intresting, but i had to wait till my daughter came last weekend to show me how to enter, i get lost with a mobile phone!, but will follow with intrest

  • norpet

    Mmm, I’m not on Twitter or Facebook. They still sound more complicated than they are worth. Guess, I’ll be sticking to the ‘ordinary’ Loquax comps for some time yet!

  • IzzySc

    Very interesting, thanks for sharing – some useful tips here!

  • shal777

    Great article. I enter and win on Twitter. The promoters want you to enter the comp as many times as possible because you are advertising to your followers, so the more you enter the more chance you have of winning and the more their name gets out in to the twitterverse. I only pick comps that have an end date and usually pick a specific couple to RT constantly and so far I have been lucky. (only small stuff like DVDs and books)

  • kevinwj

    I think as more data becomes available over a longer period of time, that these findings may alter. Either that or I’m the exception to the rule!

    I tend to enter as soon as I see the comp and so far this year have won 2x books, 2x T-shirts, £10 vouchers, 1x Magazine,1x CD and 1x DVD

    Admittedly they are only small prizes but 8 prizes in 6 weeks is not bad, bearing in mind I dip in and out during the day and only enter for things that I or my family will make use of. If I followed the “win is a win” credo, then who knows what my total results would be.

  • kevinwj

    Just to add,

    I use tweetdeck for PC and it tells me if I have already retweeted a post and doesn’t let me RT the same one twice.

  • @kevinwj are all the comps you’re winning RT to enter ones?

    What we are doing though is now picking up the app the promoter uses and the app the winner uses… see if there is any correlation.

  • silverade

    I only started tweeting properly in December and entered a RT comp 5 minutes before it ended and promptly won £50 of Debenham’s vouchers which ties in with your research. Very nice too! Since then I have got into the habit, when I switch on the laptop, of first opening Outlook Express, then Facebook, then Twitter. This means that as I plough through the emails I can swiftly enter stuff on Facebook and Twitter too. I have found Twitter to be rather addictive as I cannot resist constantly looking over to Twitter to see the most recent tweets. Since my first win I have won some shampoo & conditioner, a t-shirt and a silver heart necklace. I have won a similar number of prizes from Facebook in the same time period but they have been of higher value. I still feel like a newbie on Twitter and don’t understand about Tweetdeck etc but I’m happy with my rate of success and have encouraged my friends to join too as I think it’s interesting to be involved with this new way of entering comps which is evolving all the time. Incidentally I’m @Brackenheart on Twitter.

  • kevinwj

    @Jason It’s a mixture of RT and answer a question type, I think one win was a fastest reply but that was down to luck that I was online and had manually refreshed the timeline just at the right time. I’ve not kept track of the ratio, but would guess that most are RT and forget entries.

    Another win today, which was from an RT on the 11th.

  • kevinwj

    Just had a thought, the tweetdeck search only gives a limited number of results, so as I just did to find when I RTd my win today, I used Twitter for PC. What I noticed is that when you search it tells you who has RTd a #tag but not how many times. If promoters use the same method it would effectively negate multiple RTs and reduce it entrants only

  • Ennill1

    Don’t these companies that run ‘fastest to reply’ or ‘first three’ etc realise they are reaching out to a v small number of people that way, and generally just annoy us?

    There is a bit of a ding-dong’ going on on twitter at the moment where some users think they have identified a member using multiple identities to enter comps – think suspicion was aroused by number of wins – however it seems twitter can’t confirm that all identities are using the same computer-or-whatever-they-check.

top