From April 7- 9th, 2010, I attended both eMetrics and Search Marketing Expo (SMX Toronto). This was my first time to eMetrics Toronto but my fourth time for SMX – I was impressed as it was a really great conference for learning best practices with web analytics, SEO, SEM and social media.
I had attended many tracks and have compiled my favorite key take-aways below.
#1 Understanding the hierarchy of needs is key for web analytics strategy.
Developing a web analytics strategy requires understanding the hierarchy of needs (data, information, insights & recommendations).
John Lovett stresses that web analytics is NOT data collection and reporting, it is:
- Dependent on operational interpretation and visualization.
- Managed globally for all process and functions.
- Continual monitoring process.
When companies invest in analytics, they can create sustainable and strategic advantage.
It is easy to profess a great love for data and still fail to use that data to inform action. A lot of companies are in this category though.
John Blackmore from IBM, stresses that our challenge is to get the appropriate funding for marketing functions and you need a pretty iron clad explanation – analytics is the proof. There’s still a huge opportunity for marketers to do this better…most is not being done well.
#2: For a metrics driven culture, start with objectives.
I like how John from IBM simplified his five principles of metrics success:
- Define objective – what are you trying to do?
- Go after low hanging fruit first.
- Pick marketing battles you can win.
- No focus, no excellence.
- Test. measure, change, repeat – the true analytics culture!
#3: Web analytics is hard, you have to transform it into useful actions and it takes time.
Transforming your organization into a data driven culture is an investment in time and resources. It is easy to pull reports, but it is a lot of hard work and perseverance to have a culture where analysis can influence all company decisions.
John Lovett recommends starting with these two areas first:
- Architecting technology. The right data comes from multiple systems (web analytics, voice of customer, customer experience management, social analytics) and you must integrate data points into your process in an wholistic way.
Five (5) signs that you are doing it right with staffing web analytics:
- You have a senior person who owns it.
- Dedicated resource for web analytics.
- You know who are your analytics amateurs.
- Analytics hub supports whole company.
- Analytics hub provides insights and recommendations.
Theresa Locklear from NHL.com shared with us that her web analytics team started with just her (now the team has 4 members) and her journey to define key metrics and turning it’s organization into a data driven influence was a 2 year accomplishment. She also expressed that it doesn’t mean that they don’t have a lot more work to do…automation is one they’d love.
So how do you know if your organization is doing it right?
- Nobody says “if only we had solution x”.
- Nobody questions the money you are spending.
- Nobody uses the data because they have to, they just want to.
- Nobody is forced to tweet on twitter for help!
Developing a web analytics strategy requires understanding the hierarchy of needs (data, information, insights and recommendations).
Measurement of actionable data is a process, not a one-time event. Does your company follow a process, is there a clearly defined work-flow?
There’s six areas that you need to kick butt in
- Have a governance model in place.
- Invest in technology.
- Allocate resources to your web analytics program.
- Focus on analysis and recommendations.
- Measure impact of recommendations.
Lovett also stresses that web analytics is all about the people, process and technology – three equal measures that are important.
#4: There’s no perfect dashboard and Microsoft Excel is OK to use…for now.
When smaller organizations are getting their feet wet for the first time with analytics, there’s a misconception that the reports generated from Google Analytics are sufficient and they will tell all. But soon enough someone (and most likely a web analytics professional) will show that a custom dashboard is needed to aggregate data from many sources that Google Analytics may not have (i.e., costs from banner ads, Pay-per-click spends, data from social media and mobile, costs for press release distribution, costs for social media advertising, etc..). And this data needs to be customized to the organization goals and crystallized into more focused reports for analysis. As far as I know, there is not one dashboard solution available for small to mid sized companies.
I was lead to believe that larger websites and web teams are using a expensive dashboard applications though.
But at eMetrics I found out, that it was not the case….
I was happy and surprised when Theresa Locklear from NHL.com showed us her web analytics dashboards, which were all in Microsoft Excel. Because of the many data sources, they found excel as the best solution to do reports for key stakeholders. She showed us many samples of the dashboards they created with data, graphs and executive summaries. Some of the samples she showed us were for the 6 areas their web analytics team helps the organization with analytic driven influences (Editorial, Site Development, PR, Consumer Marketing, Social Media & Mobile, and other digital services). Her team supports 30 team club websites in addition to the main NHL.com website. I was very impressed and loved the sight of the excell sheets. Oh la la!
For my current job at ActiveState I have created ay own web analytics dashboards in Excel (consisting of a deep dive and an top level KPI dashboard) and thought I was alone developing my own homegrown dashboard – not so much now.
But it doesn’t mean that I would adopt a dashboard application that was available to me if there was one without a big price ticket. I think that this is a pain point for some web analytics managers and some companies can take advantage of this problem and develop a SaaS drag and drop dashboard.
There was repeated comments throughout the two days at eMetrics on dashboard and reporting compilation time.
- Ross Jenkins wished that he had a platform that automates mundane reporting, so that he can free up his team for hard core analysis.
- Theresa Locklear from NHL.com is wanting to automate their dashboard more, so that her team can spend more valuable time analyzing the executive summary with a more consultative approach with key stakeholders.
#5: We’re fixated too much on “what” data. Qualitative + Quantitative = Ideal.
Quantitative date is it mostly “what” like reports from analytics (i.e., visitors, click-throughs, bounce rates, sales, referrals)
Qualitative data it is mostly why, it can be consumer feedback, social networks, interviews, surveys and polls.
We should not use only one, but we should use both.
“Connecting awareness to revenue with B2B can be done with qualitative and data mining – go the final mile. ”, says Theresa Kushner from Cisco.
Sometimes you cannot close the loop to see what converted to what. Especially in BtoB long term sales. So something as simple as click to chat with some questions on how the customer has found you or asking what products they are interested in is qualitative. You can then data mine and analyze you efforts.
#6: Puking out data is not ideal, only measure what is actionable and important for you.
Avinash Kaushik, says many people fail because they don’t action on analytics – they just puke out data and reports. Context is queen and and you need to provide more context.
For example direct traffic report, look at it – it doesn’t tell you anything. What you can look at is what percentage of traffic is from free search. That is your SEO efforts and it can be worked on – it’s actionable. Direct isn’t.
Also, if you are doing a lot of online branding efforts (banner ads, PR etc.), another good way to look at your data is to segment your traffic coming in with your particular branded keyword and phrases.
This is a very good way to measure and track your branding efforts, I have recently done this for Activestate (where I work) by creating a segment and filtering all combinations (i.e., activestate, active state, komodo, komodo edit, activeperl, activetcl, activepython etc.) and then analyzing that traffic (where they are coming from, what are the high converting keywords). Again, don’t look at direct traffic – this is not accurate for gauging branding efforts.
Another way to look at traffic is to segment by campaign costs. Splice and dice data, look at what campaigns are delivering conversions.
Track social media, what are converting to orders, what are macro v.s. micro conversions.
Give more context point blank.
“Just because you can measure things, doesn’t mean you should”, is something that Theresa Kushner from Cisco shared with us. They only focus on what matters to their strategy and execution.
John Blackmore from IBM says “ If you are not going to change it, don’t measure it”. He also pointed out in one session that they didn’t report on bounce rates in the metrics because they weren’t going to act on them. We decided to focus on lower hanging fruit instead. Doesn’t mean that they won’t add it into their metric in future though.
#7: Web Analytics Certification exam ain’t easy.
Being a Web Analytics Association Member , I was fortunate to be invited to take the “Test of the Test”. The WAA association was field testing its new Web Analyst Certification Exam. This test was not easy, I received my mark and I did not get my regular A or A+ that I was accustomed to, but on the positive side, I passed.
Overall I think the test is challenging people to go beyond the tools towards analysis. Also, some of the questions are very heavy on the marketing analysis side, so I think some people from a tech /IT background may have a tough time. Even myself just completing the web analytics courses at UBC, I thought it was very hard test. Setting the bar high on this test is good for professionals who want this certification – it will mean something more than “Ya, I use Google Analytics”.
#8: Conversion optimization is all about leaving money on the table.
Khalid Saleh shows us a before and after sample of an online store page and they were tiny tweaks to fix show stoppers for some customers to convert, he points out to us and says “Never judge by how big the optimization changes are but how big the results are”.
Khalid is so correct, I have seen this personally when I used to manage BtoC online stores. One simple tweak can double the conversion rate from 1.5% to 3% (of all site visitors) and that can mean a lot when you have sales of $75,000 + per month and potential customers are on your site and can convert better.
Look for low hanging fruit within your analytics, on what should be priority for optimization. For example :
- If you have site visitors landing on your site category pages and you are getting 70% on one category and 2% on another category page, you may have an SEO issue. Investigate and fix it.
- Look at category paths, you should see clicks to level 1 to level 2 pages. If they go to child then to parent there may be a problem. They may be confused about the category.
Shari Thurow gets into site optimization v.s. SEO discussion and shares with us that we must design for people who use search engines not just search engines or just people. She shares with us 5 rules for SEO design:
- Easy navigation
- Easy to find
- Consistent layout
- Downloads quick
One of the most important things is keyword scents. Be consistent with keywords that validate what your potential customers want from the external ads to throughout their experience on your site. Have them in your primary areas (visible body copy, text at top of page). If the scent is broken they may leave.
Use your analytics to optimize your scent paths (keywords throughout your site). Thurow tells us “Sometimes you just have to show key stakeholders what kind of search terms people DON”T use.”
Khalid also shares with us that site optimization should have a road-map, figure it out first then start the optimization stages.
#9: Organizational buy in for data driven culture is key.
During one of the sessions, a panel was asked “What tool would you want that you don’t have” and John Blackmore from IMB answered “Executive by in”, the audience chuckled and related to him.
This is one of the biggest challenges for web analytic managers. It is something that we tackle in any organization and probably at different degrees.
So what can you do, to get buy in?
- Find an evangelist within your company and start there. Take some one on one time and show them opportunities that they can action on. And use this opportunity and turn it into a case study.
- Share case studies, with other team members in the organization. Tie these samples back to results (i.e., conversion rate). The ears will perk up further if they know they can contribute to this conversion and you will start to see more intelligent questions from curious team members.
Be careful when sharing reports. Have tired reports (high level, mid level and down and dirty) – don’t overwhelm executives with the down and dirty reports either. They don’t care, it is the bottom, line bigger picture for them. Know what reports to give to different team members (like a down and dirty report of banner ad campaigns/costs may be ideal for the marketing ad manager)
Do web analytics training sessions and offer that to all employees. Teach them how web analytics work, what are the data-points and make it fun. Andrea from 360training.com found some great success on educating and getting buy in from this tactic. But keep it visually stimulating and give prizes for interacting.
Never underestimate the power of free food.
#10: The future of search.
Search is not going anywhere. But as marketers you have to consider and adapt to these changes.
Web is more mobile than ever. Mobile marketing is growing. Look at India, it is growing1.2 million per month – the speed of adoption is tremendous.
Mobile search is not only on the street but on the couch while watching TV.
Search is moving from text to voice search to voice search with GPS – the merging of the apps. Think of foursquare with a GPS that is voice activated.Uncategorized | Tags: emetrics, emetrics toronto, metrics, web analytics | Comment (0)