Data. Whether you love it or hate it our world revolves around it. From personal data that identifies you, to the facts and figures in your presentation to what happens when we are navigating locations and purchasing products. Some companies have made their fortunes from the collection and aggregation of data which is then resold and reused many times over – think Nielsen, Facebook and Dunhumby. And most people would agree that the more data you have on a subject the more accurate you can predict trends, behaviours and the impact of changes. It would seem that more is better!
Why then in shopper research projects are we typically seeing samples of 8, 20 or if we are lucky 50 participants?
Of course in some cases a small sample is completely the right way to go – research should always be bespoke to the scenario in hand, not one size fits all!
One thing that typically defines a samples size is cost. Qualitative interviews, in-store intercepts and focus groups can be costly and time consuming and give us only a snapshot of our potential customer base. Eye tracking for example is incredibly powerful and allows us to see through the eyes of the customer, giving deep insight into the decision making process, search strategies and the way the consumer engages with the environment – but large samples can potentially be costly and time consuming to analyse. There is of course the argument that the cost of not having solid findings from a decent sized sample is far more expensive than increasing a research budget?
Too often research is commissioned either using the same old approaches and methods which although familiar and well used, may not be appropriate for the research. Many times questions are defined before a wider view of the objective has been determined and you could even be looking in the wrong place to even define the scope of the research. This could lead to a project that doesn’t fulfil the stakeholder brief before we even start testing just a handful of people and make potentially a catastrophic decision. Of course there are stand out cases where large samples didn’t work out – over 200,000 people did a blind taste test for “new” Coke in the mid 80’s and that didn’t work out too well for them as history played out. (Later analysis pointed to the fact that people didn’t just purchase Coke on the taste alone, but had an emotional involvement and attached some symbolic value to the original Coke product – showing that just one methodology (taste test alone) probably wasn’t the way to go, multi-discipline research is something we always promote here at Acuity.)
As well as cost, time is always a factor. The total time to collect the data, analyse the results and then implement the findings from a study can be a lengthy process, especially in the world of fast moving consumer products, seasonal trends and competitive marketplaces. The larger the sample, and the more data points we have to analyse – the less time we have of fitting into launch schedules, to meet business critical requirements and keep up with the fast paced battle at the tills. (As a separate read in the ongoing fight for consumer revenue I would recommend looking at Tim’s blog pieces on how ‘Copy Cat Brands’ leverage the branding and associations of leading brands starting with this post : http://www.acuity-intelligence.com/defending-your-brand/)
Of course even if we have all the money in the world and a near infinite amount of time to complete the research we also need to be able to capture the data in the first place. The logistics of large sample studies can be overwhelming, with personnel, recruitment, travel costs, store permissions and equipment availability are all factors in the overall equation. So with all this considered it is sometimes obvious why larger sample studies aren’t undertaken maybe as often as they should be.
With the rapid evolution of sensor based technology, increased computing power, cloud networks and 4G/LTE connectivity all of a sudden we do have the tools to do mass sample studies, and make them easily accessible.
Most of our clients in the FMCG or retail space face the challenge to understand the complete path to purchase and to break the shopper journey down into different critical moments like in-home and out of-home interactions with the brand and product, an individual’s journey throughout a whole store, measuring how shoppers behave in a specific category, how POS materials affect their decision making or even identifying missed sales opportunities on a shelf. We at Acuity know that not all these questions can be answered by just one tool, which is why we look beyond eye tracking and provide our clients with a wide range of powerful tools that help you understand those critical moments.
From the advent of marketing profiling potential customers by their gender or age has been seen as key to success, and rightly so. However how much of a disconnect is there between the people we present the thousands of brand impressions at compared to those browsing and shopping in store? Measuring age and gender (and also emotion in some cases) at touch points around the store can help retailers and brands determine if their assumptions about their target markets are correct, if they maybe need to cross an age or gender gap with their messaging or in fact – if they have simply got it all wrong! Using a technology such as Quividi (http://www.acuity-ets.com/product/quividi/) and a discreet camera we can measure not only the age and gender of shoppers but also their dwell time and attention to the shelf or media while capturing basic emotional responses as well. We can take it a step further and change displayed digital media relative to the audience in real time – so while the system is collecting the number of observers it can also deliver tailored marketing messages to drive traffic to a product, or educate (or entertain) customers… have a look at Quividi in action with this interactive billboard (and in this case they didn’t do a good job of making the camera discreet!)….
The size of the sample we capture with Quividi of course depends on the traffic in the environment we are testing, but the system can sit there 24/7 observing and capturing the information, uploading it regularly to a cloud based server – so it ticks all of the boxes for rapid deplyment, ease of use and quick access to large data samples. The video feed is processed live and the viewer categorised, and this information can be available with just 15 minutes delay. The product is typically a simple install just requiring a processing unit and camera to be powered and tucked away discreetly – for example in a digital sign, behind a fixture or floor stand or behind digital signage.
If you want to understand how people move around a space to identify traffic flow, choke points and areas where people dwell we typically widen our capture area to either an aisle, category or store level environment. And to properly understand trends we also need to have a large data set to allow us to measure behaviour at different times, dates and with seasonal influences. Utilising a tool such as uCount (http://www.acuity-ets.com/product/ucount/) we use either a wall or ceiling mounted sensor to measure footfall in an area and can then create areas of interest (such as promotional areas, checkouts or service points) to analyse patterns of behaviour, typical journeys, excessive delays in way finding and peak traffic times. Everything is processed on the device and uploaded to a cloud based system and you can access the data on a hourly basis, meaning iterative changes can be applied and observed with little delay with sample sizes in the hundreds or thousands! There is no need to carry a token or fob, there is no Bluetooth or Wi-Fi tracking and the flexibility of the camera mounting options open up testing across many locations, and the set-up is incredibly simple. Some key example use cases would be pop up outlets, promotional areas in stores or foyer entrances in shopping mall spaces.
For more granular detail without using multiple camera technology to cover a larger area you can utilise Shoppermotion (http://www.acuity-ets.com/product/shoppermotion/). This is a world leading solution currently used by some of the largest names in retail. Where customers use a bag, trolley or basket Shoppermotion low cost, discreet tokens track journeys and paths with an accuracy of around 50cm, 3-4x more accurate that Wi-Fi tracking (Consider that when locating a shopper 2m of error could be the difference between placing a consumer in the health and beauty aisle or frozen foods!). A simple network of sensors keep the installation to a minimum and provide near real-time reporting of behaviour with no requirements for beacons on every aisle, Wi-Fi or Bluetooth to be monitored on shoppers phones or the need for a mobile app. The software’s open architecture allows integration of other data streams (EPOS, Quividi, queue management systems, etc.) to feed into the dashboard allowing KPI’s to be visualised, real time queue management, statistics over time and much more. From identifying which routes around the store are more common to looking at dwell times across categories, fixtures and bays the Shoppermotion platform is incredibly powerful while making the data easily accessible through a web browser interface. While a slightly more complex installation than uCount, Shoppermotion gives you more detailed data across entire retail spaces with high levels of accuracy. Combine this with the speed of access to the data in the user interface and it creates a very powerful research, analysis and store management platform.
Understanding who is shopping our store, and how they are moving around the store is fantastic information for retailers, brands and FMCG’s but often they really want to understand what is happening at the fixture and where loyalty cards and EPOS data can give us an insight into what is purchased there is as much value, if not more, in knowing what wasn’t purchased, how complex the buying process was and were other brands or products considered before the shopper made a decision. Shopperception (http://www.acuity-ets.com/product/shopperception/) is an innovative platform that allows us to do just that. Monitoring shopper traffic, dwell time and interactions at a SKU level allows us to really understand how a product made it into the basket or didn’t. The dashboard allows us to drill into the number of pickups, touches and returns on a brand, category or product, it lets us measure lost opportunities to competitor products and see what items were considered during the purchase decision. We can measure the conversion funnel from end to end, identify sales performance against share of shelf, measure basket split by brand and much more – all without the requirement for observers in store with data processing automated and scalable from a single bay to a category or beyond. The online analysis tools allow the data to be polled for a wide range of queries and the typical data set we see on projects are in the region of several thousand customers, spanning days, weeks or months. While set-up is a little more complicated than other technologies the granularity and detail of the data is unsurpassed and with statistics being available next day insights can be collected quickly and with the confidence that findings are being formed on large participant samples.
While in some cases these tools alone don’t give you all the answers they do allow us to identify if we understand where issues or potential performance increases might happen, they help us identify what our next steps are or maybe they reassure us that recent changes have been effective. They also give us large databases of benchmarks to work from, which can only be a positive benefit. Other products or methods allow us to dive further into this data to gather more insights, and combining tool and approaches can multiply the benefits of a research project.
At Acuity we don’t just do eye tracking and hopefully (after a 9 year wait….) our new website reflects that and you can see many of the other products and tools we supply, and we constantly evaluating new breakthrough technologies to help deliver bigger, better and more valuable insights into human behaviour. Get in touch and challenge us with your problems and we will be happy to discus and advise you on ways to help you get the data you need to answer your research questions.