The sample is the term used to describe those people that participate in the research. For clarity we use the term sample-frame to describe the group of people invited to participate and final-sample to refer to the sub-group that actually went on to do some or all of the survey.

For some studies the sample-frame is drawn from lists and databases already in possession of the researcher. The HR department can provide a list of names for a staff survey and similarly the sales and marketing team should have handy lists of customers and prospects for a customer or market intelligence study. Care should be taken that you have the legal right to contact people in the sample-frame and that doing so won’t place you in breach of the Data Protection Act.

Where things get a little more interesting (and expensive) is when sample has to be recruited for the study. Typical studies requiring external sample would include consumer studies and business-to-business research. We see three categories of external sample provision for online surveys:

• Ad based recruitment. In this scenario you place adverts for your survey and then hope people read and respond to them. Response rates for this method are incredibly low, don’t expect more than fractions of a percent and therefore you will need them to be seen by many thousands of people for them to have any impact. Online ads have the benefit that you can measure the response and see what works for you whereas printed ads are a little more ‘finger in the air’.

  If you do go down this route then Google Ads is not a bad place to start and something we’ve used before at ORS. Its easy to use and you can pay per-click on your advert and set the budget you want to spend.

  One action you can take with printed ads is to register a domain name and point that at your survey. This means people can use www.thebigshopsurvey.co.uk rather than a typically long and complicated survey URL full of ?:// and other easily forgetable symbols.

• List brokers. List brokers will send your message out to a list of email or postal addresses for a fee, there is no incentive for the recipient offered by the broker so you must ensure that you are offering one and your communication is well designed (see below). In our experience the response rate using this method is low at around 1% or less.

  In my opinion this method is best avoided as the quality and often the legality of the lists concerned is somewhat questionable. Often people on these lists know nothing about them and their details have been bought from other companies or farmed online by software programs.

• Panel companies/ Reward programs. These companies generally offer the best source of external sample for online research. The program members have signed up and are actively involved. They earn a points based reward for taking surveys which they can cash in for goods or money when they have enough. That means the incentivisation aspects are taken care of for you as well. You will find yourself paying per-completion which does mean that either the job gets done or the provider goes without full payment. The down-side is that this can be an expensive exercise. I advise obtaining quotes from at least two or three suppliers and ensure that you are meticulous in your specifications to them.

  ORS has lots of experience of obtaining quotes and brokering deals with sample providers. In the past we have obtained quotes for quality sample at £3 per respondent from provider A and £18 from provider B for the same job. Multiply that by 1,000 and you can appreciate the importance of getting this right!

  Before getting a quote the screening and quota criteria must be defined precisely. If a panel company is asked for a cost for 1,000 car buyers they will quote price £x. When it comes to the actual survey and it transpires that there is a screening criteria that specifies respondents must have bought a car within the last 6 months and there are quotas that specify 250 respondents each in North, South, East and West and that they should be 50/50 Male/Female the £x no longer works. It then becomes £y which is about 4 times as expensive. Even worse, they might not be able to do it all. This is a position best avoided, particularly if you are a research agency that’s already quoted the £x price to your client.

  Another aspect to consider when using external sample is how to integrate the panel companies system with the survey system being used. This is often a draw back with using the DIY survey tools on the market.

  ORS can look after this for you and even allow mixed sources where appropriate, this service is included in the ORS Standard Survey Package.