Having spent most of my day in Google Analytics I found myself discussing data analysis with a friend over a pint this evening. They were unhappy with the current state of their data set, and had a predefined set of issues which were seeminly dead-end issues. I wanted to know more so noted down the six issues that they mentioned. After a few minutes blurting out as much data-related information as I could muster we’d managed to solve some of the key roadblocks that had been holding back genuine progression.
Here’s six of the myths that had been holding back any data-driven insight and a summary of my response.
Myth: Historic data contains all the information required to produce meaningful insight.
Whilst historic is undoubtably a very important element of data analysis, real-time data is becoming increasingly important, too. Real-time gives minute-by-minute insight into campaign performance and consumer behaviour which is often lost when viewing data over longer periods.
Myth: Data analysis is all about numbers: Analysing individual behaviour is pointless.
Although analysing the actions of a single individual may be a case of ‘sample of one’, there’s several good reasons why analysing behaviour in a qualitative (rather than quantitive) nature may be beneficial, for example testing user journey or expereince analysis.
Myth: “The data is really insightful, but I already knew that – I know my target audience”.
The reality is that data driven decision-making is much more assured, accurate and financially secure than making assumptions about an audience or experience. Whilst there may be a good understanding of audiences, being informed with strong data analysis is essensial.
Myth: Benchmarks vary too much from site to site, and they’re too case specific.
It’s true that there’s an abudence of benchmarking sites across the internet, each quoting different statistics to aim towards. However most reputable sources will give detailed explenations of sample data: It’s just a case of finding the most suitable set of data. Plus there’s always your own (year on year) data to benchmark against if you can’t find the statistics you’re after through third-parties.
Myth: Every company has data quality issues, I’m no different to other organisations.
It’s fair to say that data quality can vary massively between company, but in my experience as a general rule (with one notable exception) company data has been decent. If you find yourself looking at thousands of rows of questionable data there are several quick wins which can be automated in excel (deduplication, activity filtering, behavioural modelling) which don’t require expensive data analysis teams.
Myth: It’s impossible to forecast accurate figures as the market is too volitile!
Producing a solid and accurate forecast can be tricky, but there’s really no excuse for not being able to give a rough idea of performance over a given time. YoY data coupled with third-party analytics and variable profiling is often enough to make informed decisions based on available data. For perticularly rocky marketplaces you may need further data sets such as similar markets and market predictions.
Hopefully the six points above demonstrate the freedom and insight that data can provide. Wether it’s a quality issue, granularity concern or measurement problem, data has a natural self-remedial ability given the right triggers. If you’ve got a data-related question drop it in the comments below and I’ll do my best to answer!