Ignorance is not bliss

by Marketing ON 20 May 2010

As a marketing manager (we do most of our stuff online) I have a budget to look after and my main goal is to maximise the return from my marketing investments. I have lots of tactical choices available – banner ads, search, print, emails, newsletter sponsorship, social media and so forth. To decide where to allocate my hard-earned budget, I generally look to the history of my various campaigns and spend more where I know I will get a better return.

The hidden problem for marketers is that many of the measurement tools being used to determine campaign outcomes are totally bogus and so the results being used to determine future marketing investments are false. Truly, you could be operating in a vacuum of blissful ignorance and not even know it!

Most of todays online web analytics tools, including Google Analytics, Omniture, Webtrends and Yahoo Analytics default to what is known as last click attribution – that is, they count the value of the conversion (sale, booking event etc) towards the last marketing activity that happened. Last Click attribution is good for measuring conversion events but will not tell you which campaign brought the customer to you in the first place – for that you would measure the First Click. And what about the other campaigns along the way that influenced the conversion? How do you measure these and make sure they get some credit as well?

The key takeaway is that depending on the attribution model you are using, your return on investment (ROI) and return on ad spend (ROAS) may vary dramatically, and may lead you to make different (and possibly very wrong) decisions about how to best optimise your marketing efforts.

Whilst there are some very advanced attribution techniques being deployed by some companies, the general consensus today is that an approach looking at First, Last and Average Click, side by side, is a sensible and achievable approach – and it can be done with minimal effort (and possibly a spreadsheet, if you must). Such an approach will show you which campaigns are working better for acquisition, influence and conversion – you can then weight the results to drive the campaign mix you are looking for.

A recommended action plan:

Find out which attribution model you are currently using,
Re-assess your current marketing program now that you know what your are really measuring,
Implement a First, Last and Average Click attribution model for your business,
Reconsider your mix strategy for acquisition, influence and conversion,
Make more informed and effective investment decisions,
Bask in the glory of knowing you have improved your marketing effectiveness, and
Spend your ‘Management By Objective’ bonus!

This is my personal blog. The views expressed here are my own and do not represent those of my employer, Coremetrics.

Australian Business Community versus Free Web Analytics

by Marketing ON 24 November 2009

There was a gasp from the gallery! The result was not at all what the industry pundits had expected. In the case of the Australian Business Community versus Free Web Analytics, the jury had weighed up the pros and cons, considered the evidence and made their surprising decision: Free Web Analytics was guilty as charged of the crime of grossly overstating the benefit of its services and leading Australian businesses astray.

 In his judgement, the Chief Justice of the High Online Court explained that the common perception that Google Analytics was the most widely used measuring vehicle for online performance for web sites in Australia simply did not stand up to a scrutiny of the facts. “This miscarriage of justice is a travesty and should be exposed for all to see. Sure, many companies use Google Analytics, and why shouldn’t they?” he said. “After all, it is presented as a ‘free’ service, but the fact is that the majority of leading web sites in this country actually use paid analytics services. Businesses need to know this otherwise they are basing major decisions on incorrect assumptions.”

 In presenting the case for the Australian Business Community, we reviewed a group of websites crowned as ‘leaders’ by two of our nation’s most recent studies: The Amber Awards, The Australian Interactive Media Industry Association Awards, and the HitWise Online Performance Awards presented on 24 March 2009. Of all the Amber Award winners, more than 2/3 (66%) use a paid analytics service, while the Hitwise Awards showed that more than 50% of the winners using a paid analytics service.

 Not surprising is that many of these leading web sites also use Google Analytics, as well as a paid analytics service. Furthermore, the most commonly used paid analytics services were the top recognised international vendors (in alphabetical order) Coremetrics, Omniture and WebTrends. 

 A closer look at the award winning web sites revealed a trend that, in general, the more creative sites and those primarily outsourced to web development and advertising companies had a higher tendency to use the free services, while the websites that were maintained in-house were overseen by tightly managed and dedicated web marketing teams.   

This leads us to two questions: 

Why are the top sites paying for web analytics? 
Who is perpetuating the myth that Google Analytics is all a good business needs?

Companies pay for web analytics for a myriad of reasons but the most commonly heard include: 

The depth of reporting from free tools does not match the more advanced paid services
The ownership of the data is important – larger and more successful organisations tend to place a real value on owning the information about their marketing channels 
The need for information reliability and security, which generally comes from a formal Service Level Agreement, can only be provided by a paid service 
Technical support and training from the vendor is important 
Free software does not equal $0 total cost of ownership, and 
Analytics can be complex and many leading companies are looking for a business partnership, not just a vendor.

 So, who benefits from promoting the simplicity of a ‘free’ service? The vendors like Google provide analytics as a tool to help promote the expansion of their primary product (because a common outcome is that you buy more Adwords), as do Yahoo!, who also have a ‘free’ service to help promote their online services. In addition to these guys, the main promoters of the free services seem to be advertising agencies and web developers. For many agencies it an easy way to create the impression that they’re measuring the results of campaigns and can also be part of their revenue stream. For web developers, many of whom know no better, it is a quick solution for their customer needs.   

What to learn from this landmark case? The primary takeaway is that most of the major and more successful websites in Australia use paid analytics tools to drive programs of continuous improvement through their online marketing programs. Free tools are a good place to start but the really successful players value their results and invest in solutions to measure their growth and drive change.    

“This is my personal blog. The views expressed here are my own and do not represent those of my employer, Coremetrics.”

Is your shopping cart only half full?

by Marketing ON 8 January 2009

The interest in the recent Online Retailer event held in Sydney highlighted that most companies appear to be still struggling to fully benefit from ecommerce and offer their clients/customers a new convenient sales channel.

A number of Australian retailers are working to refine their online strategies and understand that if they are successful they can potentially make tens of millions more in revenue. With many Australian retailers still learning the ropes, you competition may only be a few steps ahead and it’s not too late to catch up.

The common project questions are what platform, shopping cart and web analytics will be best suited to your industry and budget. What can often confuse business is the sheer number of potential store platforms and shopping carts available.

For simplicity we have listed seven of the most common solutions in typically order of cost:

Paypal
Amazon/eBay stores
Out-of-the-box hosted
Custom built by web developer
Open sourced solutions
Proprietary solutions, and
Enterprise solutions.

The store cost also varies from a small investment in time and possible ongoing charge per sale to enterprise solutions costing millions. The first three options are the cheapest and quickest solutions but are not suitable for competitive industries or large stores. Many custom built solutions by web developers while appearing cost effective can often have longer term limitations, can be inflexible, non-search friendly and make it difficult to track ROI.

The best solution for ecommerce stores is typically an open sourced solution, as they offer the most flexibility, support, plugins and customisations. Proprietary solutions can be suited to complex projects or need implementation within existing CMS platforms, they can prove expensive and not always the best solution. The enterprise solution is the type that Walmart, Amazon or Woolworths would select and can be a combination of proprietary and open sourced solutions.

The main factor that can rapidly increase costs of an online store is the addition of multiple solutions such as:

Back-end fulfillment
Customer relationship management
Email marketing integration
Geo targeting/personalisation
Multiple stores
Integration with custom applications
Onsite search engines, and
Usability testing and implementation.

The more advanced options can even replace other existing sales channels as they can offer more potential benefits allowing visitors to receive the best possible experience and better ROI for the business. While these can increase initial launch costs the benefit of many of these solutions can provide to your online store with better conversion rates, reduced processing costs and even increase visitor traffic.

In this current economy many projects don’t always have the budget for a complete website rebuild so there are two effective solutions to increase your ROI. If you have an existing website solutions such as SLI Systems offer intelligent search solutions which can improve conversion rates and can assist with improved search optimisation campaign results. There are also web usability firms such as Peak Usability that have advised my clients that typical conversion rates should be around 2% and websites should be aiming for around 5%.

The focus in a successful ecommerce business should always be on quality foundations supported by the ability to correctly measure ROI and success. The key to better understanding your visitors and behaviours is having correctly implemented web analytics running within your shopping cart that combines visitor data and revenue.

Around one third of websites are still not using web analytics software and rely on the basic web server logs to understand the success of their online marketing campaigns. The problem that many of these web server logs while free, do not offer business the necessary information required to track their campaigns or refine their online marketing spend.

If you are driving visitors to your online store you need correctly implemented web analytics so you can track how much revenue each campaign or medium delivers the best revenue. While most web analytics packages support campaign tracking some allow for advanced integration with email, video & social media platforms.

Most people understand that the leading web analytics solution is now Google Analytics based purely on the volume of websites, across all industries followed by Omniture. According to my extensive research of the world’s top 1,200 public companies a majority of the world’s largest retailers are split between Google Analytics and Omniture.

While enterprise solutions such as Omniture are often suitable for large ecommerce websites, its setup and costs can make it harder for business to justify. There are other web analytics solutions that fit between Omniture and Google Analytics, listed below with links to their clients/case studies.

If you are in a particular industry it can be a good guide to check the top few sites and follow their solution for web analytics as customised packages maybe available for particular industries. So using web usability testing, onsite search and web analytics you can measure how many visitors have their shopping cart half full and better help them fill it.

Case study success stories: