Sunday, December 30, 2012

Marketo review after one year of using it

It’s been one year since we moved from Eloqua 9 to Marketo and this is an update on our progress.  Overall, I’m satisfied with what we’ve been able to accomplish with the combination of Marketo and  We’re a start-up with limited resources and we’ve been able to develop some fairly sophisticated tracking, lead scoring, lead nurturing, and triggered alerts to drive sales and marketing processes.  Other than needing a little help from a contract web developer for two items, the marketing team has been able to do everything ourselves using Marketo and
Here are some of our key accomplishments using Marketo and
·         We built a robust tracking system that allows us to identify the source of leads and associate that information with the campaigns that the leads interact with.  We built a set of javascript cookies that allow us to capture tracking codes in URL parameters and store them in the lead’s browser for 30 days.  The cookies work on both Marketo and non-marketo web pages.  This lets us understand where our leads are coming from and which investments and campaigns are the most effective.  Ultimately, this gives us the ability to develop cost per lead and cost per opportunity metrics.  In the coming months, we’re planning to expand our URL parameters and cookie tracking capabilities to capture more insights and to enable more complex testing scenarios.  For example, to be able to test which link positions in an email (pre-header, banner ad, email body, call to action) drive the most clicks, one would need a separate tracking code (i.e. url parameter) to be able to measure that.
·         We built lead scoring models using both Marketo and  We have a sales qualification scoring model that was built using formulas and we’re getting ready to deploy a demographic (i.e. fit ) scoring model based on Marketo smart lists and smart campaigns.  In the coming months, we’re planning to add behavior lead scoring models.  Probably the most compelling part of developing the lead scoring model was the group exercise that was used to develop the sales qualification model.  The group process helped get everyone on the same page in terms of what a qualified lead is – and that eliminates/reduces the discussion between marketing, inside sales, and field sales on lead quality.
·         We launched a series of reports and real-time alerts that are based on lead behavior and downloads to enable inside sales to follow up quickly on lead activity.  Research shows that you can significantly increase sales’ ability to connect with a lead if sales attempts to contact that lead within a short time following lead activity (e.g. when a lead downloads a whitepaper, then try to follow up within 5 minutes).  We’re in the process of testing those research findings as they apply to our business to improve our ability to connect with leads.  On the alerts, we’ve found that in some cases, Marketo is a better vehicle for initiating an alert because it doesn’t distinguish between a lead and a contact in  Whereas in, you always have to develop separate alerts for leads and contacts.  One of the challenges we’re facing is trying to measure the time it takes inside sales to respond to lead activity.  The plan is to develop a apex trigger to be able to identify the first sales follow up activity that occurs after a lead action occurs.
·         In email marketing, we use Marketo for sending mass emails, email nurturing sequences, and personalized emails (e.g. Dear Fname, From: Lead Owner).  One of the painful lessons we learned was to make sure your database is clean before you start sending personalized emails – in terms of Lead Owner names, job titles, email addresses, and phone numbers.  In the coming months, we plan test using triggered emails as a way to improve conversion and connect rates (i.e. emails sent automatically in response to lead activity).  In general, we haven’t seen a lift in conversion rates by using nurturing sequences versus our control email treatment.  We tested email nurturing sequences by vertical and geographic region.  We plan to continue testing nurturing sequences and triggered emails.
·         On our landing pages we’ve deployed lead scoring questions using progressive profiling.  If you’re not familiar with progressive profiling, it lets you expose leads to different questions with each incremental registration.  This lets you get more data in the registration process without creating forms that are so long that they would impact conversion rates.  Using progressive profiling, we’re able to complete our sales qual score model questions after two downloads.  In the first registration, a lead gets 3 out of 5 qual score questions and the remaining 2 qual score questions on the second registration.
·         On the side, we built a series of analytic snapshots that let us see trend data.  We created analytic snapshots to see how the lead funnel and opportunity funnel are changing over time.  At the present time, this is more of a management tool than a forecasting tool.
Going back to the original premise of this blog, evaluating marketing automation packages, here are some of my recommendations:
·         Consider your level of expertise using marketing automation and CRM solutions.  Then consider both your short-term and long-term goals for these tools to try to understand what the total cost of ownership will be to achieve those goals.  For example, will you be able to do it yourself or will you have to hire an external consultant every time to want to expand the functionality that you’re using in the marketing automation package?
·         Marketo is a great platform for some who wants to "do it yourself."  You can do sophisticated marketing automation and tracking without requiring a dedicated admin with specialized programming skills.  I found that the “fast” (e.g. free) implementation was moderately effective and the key to our success using Marketo was premiere support.  You can read my previous posts, but the value of Marketo premiere support was that they helped us learn how to use Marketo to solve our specific business needs.  You get a named support agent and fast response time.  It’s an affordable solution and I would highly recommend Marketo premiere support for at least the first year.  In terms of implementation time, we were using Eloqua 9 and it took us two months to get everything transitioned and running on Marketo before we were ready to shut off Eloqua.  However, you have to keep in mind that we were doing the Marketo implementation in our spare time when we weren't developing and executing lead generation programs.
·         Eloqua is a powerful package, but almost everything requires custom javascript programming to be able to implement it.  For example, to capture a URL parameter using Query Strings in Eloqua requires custom javascript on the landing page.  In Marketo, configuring a registration form to capture URL parameters is a simple picklist process.  In my opinion, Eloqua requires a dedicated system administrator with javascript programming expertise to be able to leverage the platform (or a big budget to be able to fund that through consultants).  If you’re looking at Eloqua, I encourage you to consider the cost and time to deploy it initially and then try to plan for what happens after the consultants leave and you want to change something or take advantage of another feature – try to evaluate the total cost of ownership from both a budget and expertise perspective.
·         Act-On is an easy-to-use and affordable marketing automation platform with basic functionality.  It has a simple user interface and you can get started very quickly.  Whereas the implementation and training with Eloqua and Marketo takes weeks.  If you have limited resources, expertise, and goals for marketing automation, then Act-On is a great solution.  For example, if I just wanted to: send emails, create basic registration pages, integrate registration with a webinar platform, and implement a basic lead scoring model, then Act-On is a great solution.
For all three marketing automation packages that we looked at the analytics are lacking.  I still do most analysis by exporting data from Marketo and into excel.  All three packages have limited abilities to segment groups of data by activity and data ranges.  Previous blog posts describe the challenges that I’ve faced in trying to analyze engagement levels to create different segment treatments based on engagement level.  For example, none of the packages made it easy to identify all leads that have clicked just zero, one, and two times in the past 90 days. 
In Marketo’s case, the analytics are designed for looking at an individual lead’s behavior instead of groups of leads’ behavior.  Marketo captures every activity that a lead does in a lead activity log.  The activity log is great for discovering and debugging what an individual lead has done.  However, you can only export activity logs for one individual at time.  That means if you wanted to analyze the activity logs of everyone that converted into a won opportunity, you would have to export each individual activity log and then combine them all into a single dataset to do that analysis.  If Marketo could make it possible to export activity logs for segments (i.e. smart lists), then it would be a much more powerful platform for analysis.
Overall, after one year, I’m happy with the choice that we made to use Marketo and I would choose it again given the resource constraints that a start-up works within.  Moving forward, I’ll try to continue to post interesting things that we develop and deploy in Marketo and