Making Good Marketing Decisions despite of faulty Analytics data

 

Google Analytics (GA) data is not very accurate and sometimes could lead to dangerous conclusions if you are not very careful. Don’t get me wrong, I am a big fan of Google Analytics. In fact I am also a two times GAIQ certified for this very reason. But at the same time I don’t want to keep you in the dark.

If you are a GA premium user like me, you understand this pain more than anyone else.  I am not the first one and certainly not the only one who has raised issues about data inaccuracies in GA.

Both Crazy Egg and Blastam have raised such issues in the past through fantastic posts which are well worth a read:

  1. Why Is Google Analytics Inaccurate?
  2. Can You Trust Your Google Analytics Data?

 

I also wrote three posts which talk about fixing particular data collection issues in GA:

  1. Calculating True Conversion Rate in Google Analytics
  2. Adjusting Bounce Rate by Calculating Time spent on the Page
  3. Google Analytics Data Sampling – Complete Guide

I just want you to be aware of the GA issues before we move forward.  In this post I won’t talk about the current data collection issues in GA. My focus will be on how to work around these issues and make the best use of whatever data you have got.

 

Perfection is the enemy of Good

seeking-perfection

 It took me a long time to understand that the life doesn’t need to be perfect before you can be happy or you need perfect data before you can take perfect marketing decisions. If you seek perfection then majority of time you will find yourself unhappy because something goes perfect only once in a while.

Another downside of seeking perfection is ‘procrastination’.

If you spend majority of your time in trying to collect the perfect data so that you can take that perfect business decision then there is a high probability that at the end of the day you won’t take any decision/action and taking timely decisions is so important in today’s cut throat competition. Moreover an imperfect decision is always better than no decision.

Above all, no analytics tool is perfect. You can’t expect 100% accurate data from any analytics tool out there and GA is no exception. So avoid being obsessed about collecting the perfect data and be happy with the good enough data : )

 

Understand the business and get the “Context”

You should always start your analysis assuming that you have no access to your client’s GA account. Now how you will optimize the website for conversions? Well, it is quite simple.

Browse the client’s website and ask tons of questions. You don’t need GA or any other analytics tool to develop great understanding of the client’s business.

There is a common misconception that you can develop great understanding of a business just by diving into the analytics reports. This is simply not true. GA reports are just huge collection of website usage data. They can’t spell out any insight on their own to you.

If you want to gain insight from your GA reports then you need to know the context beforehand. This is the context in which you will eventually collect, analyze and interpret the GA data or any data and take business decisions.

 

To know the context you need great understanding of the business beforehand. And you get this great understanding by browsing the client’s website, using all of the sites features and asking your client tons of questions about his business, not by diving into GA reports.

80% of your analytics problems are solved even before you look into your first GA report once you have developed that great understanding.

Once you get the right context, you will interpret the data correctly and you will take good business decisions regardless of data collection, data integration, data sampling and other analytics issues.

Since you already know the context beforehand, if something is not right with your analytics data you will know that immediately. You will say to yourself, it can’t be possible.

For example if you already know that July is a peak season for your client and still the analytics reports are showing less than average sales then it means something could be wrong with ecommerce tracking or the website itself, may be page load time has increased?

Similarly, you won’t go into the panic mode when you see that traffic to the website has gone down drastically despite your tremendous SEO efforts simply because you know beforehand that you have now entered into the off peak season and consequently demand for your product has gone down.

In fact you will always be in a better position if you develop your business understanding without using GA reports and then later align your understanding with the insight you get from GA reports to determine data discrepancies and other analytics issues.

This is in fact a fire shot way to find any issue with your analytics reports in just first glance.

 

Use at least two Analytics Tools

If you are using two or more analytics tools to gain insight then your probability of interpreting the data and taking the right business decision increases by several folds.

For example if Google Analytics reports to you that your sales have increased by 20% in the last one month and Omniture reports to you that your sales have increased by 30% in the last one month then one thing is certain and that is your sales have actually increased in the last one month.

If you choose to use only one analytics tool then you can never be 100% sure about your increase in sales.

Similarly if Google Analytics reports to you that your sales have increased by 20% in the last one month and Omniture reports to you that your sales have decreased by 30% in the last one month then one thing is certain and that is one of the tools is collecting and reporting inaccurate data.

If you choose to use only one analytics tool then you may never be able to find such data collection issues.

Note: No two analytics tools report same website usage data for one website. It is quite common and normal. So focus on trends instead of the actual numbers.

 

Segment the data

No matter how bad your data is, not segmenting it will only make it worse.

Moreover segmentation is the key to successful optimization and taking the right business decisions.

You need to segment the data to its most granular level before you interpret it or take any decision.

This is because in aggregate form you will never truly get the real insight.

For example the Goal Conversion Rate of a website can be very misleading because it takes into account every visit happened on your website from all over the world. Not every visit can lead to conversion and certainly not the visits from geo locations which are not your target market.

So in order to get better insight you need to segment your goal conversion rate. Talk about like Goal conversion rate of the branded organic search traffic in your target market area (like London) and then take decisions on the basis of such insight.

If you need more convincing then check out this post from Avinash which explains the importance of data segmentation much better: Web Analytics Segmentation: Do Or Die, There Is No Try!

 

Get your Maths and Stats right

Google Analytics reports are full of averages and if you don’t know how averages work then you can easily misinterpret them.

One of the most misunderstood ratio metrics is ‘conversion rate’.

Because of poor stats skills, many marketers have no idea that conversion rate can negatively correlates with sales and profit.

They think that the conversion rate always positively correlates with conversions i.e. as conversion rate increases, sales always increases and so is profit. But this is not true. More about it is in this post: Case Study: Why you should Stop Optimizing for Conversion Rate

Similarly you can not double your sales just by doubling your marketing budget. It doesn’t work that way.

If campaign ‘A’ conversion rate is 10% and campaign ‘B’ conversion rate is 20%, then it does not always mean that campaign ‘B’ is performing better than campaign ‘A’. You first need to make sure that the difference in conversion rates is statistical significant before you can take any decision.

Similarly, when your website conversion rate jumps from 10% to 12% then it is not 2% rise in conversion rate.

Related Post: Beginners Guide to Maths and Stats behind Web Analytics 

Poor understanding of maths and stats is the fire shot way to failure in any analysis even with the most accurate data in hand.

No matter how good or bad your analytics data is, not using the correct maths and stats will always make it worse.

 

Look at the big picture and not the raw numbers

You know it by now that GA data is not very accurate and its metrics could be 10 to 80% off the mark depending upon your traffic size.

You can easily change the value of almost any metric by just changing your data sample size. You also know that no two analytics tools report same website usage data even for one website.

Because of these limitations, it is not wise to rely on raw numbers or any single metric for data interpretation. You need to focus on the bigger picture and that is the trend.

Measure the performance of your campaigns at the product or page level instead of the keyword level.  Don’t get bogged down into minute details: “oh GA is reporting 300 visits but Omniture is reporting only 210 visits’ or this keyword generated $200 but last time it generated $500”

Such type of analysis will just keep you busy in focusing on the 80% that doesn’t really matter instead of the 20% that really matters. Yes I am talking about the mighty 80/20 rule here.

80% of your output comes from 20% of your input. 80% of your sales come from 20% of the products. So you need to find that 20% and just work on it relentlessly.

Other Posts you may find useful:

 

 

About the Author:



My business thrives on referrals, so I really appreciate recommendations to people who would benefit from my help.Please feel free to endorse/forward my LinkedIn Profile to your clients, colleagues, friends and others you feel would benefit from SEO, PPC or Web Analytics.