Social media strategy and ROI ... again!

A recent eMarketer study highlighted that an ROI driven social media strategy was needed before businesses can actually start to invest in and reap the rewards. It was pointed, based on the 2 facts below, that social media users are not seeing the profits as anticipated and that there was a lack of data to support investments:


*  Only 35% are reported to have profited from social media through increased leads – these are also the ones who would invest in a social strategy and have staff dedicated to analyzing social media efforts.
*  The biggest hurdle to social media strategy is the lack of data to measure the ROI and a subsequent executive buy-in for greater investments – nearly 60% cited these as primary reasons for implementing social media strategy.

These are telling facts!

I maintained in this blog that measuring social media can be challenging purely in terms of an ROI model (net of revenue and cost), simply because the scale is not there for a typical business (in my experience). Instead, we need to utilize social media as a means to other insights (users and product) that may contribute to improving ROI through more traditional media – ones where we can easily setup a model to quantify revenue and cost for a net return (my earlier post on this topic:  http://www.analyticsheaven.com/2010/03/measuring-social-media-user-vs-product.html). 


I believe the social media efforts need to be looked at as pure investment into the future. The immediate benefits can be had from sampling and testing approach I proposed in the above post. Let’s look at social media data to learn more about our products, consumers and competitors, so we can make better decisions about our current marketing efforts. Hint - the consumers may tweet about a certain product feature they don’t like; certain types/demographics of consumers may be more interested in the brand/product; there may be a buzz about the competitive offerings that may need more attention; and so on.

Building the intelligence model from social media may be a simpler way to look at the ROI than trying to build an ROI model which, as the report highlighted, may not be easily done due to lack of data and appropriate mathematical model. 


Other insights, ideas ... please share!

Higher marketing ROI with traffic source contribution

Users find multiple ways to come to our website and represent different interests. But if we are treating them equally, then we are doing ourselves a disservice by not capturing the optimum value of each of our traffic sources. I have talked about targeted landing page experience, in this blog, for traffic from search, navigation, etc., as a means to improving stickiness and engagement. However, to please our finance folks, we also need to assign a value – an ROI or contribution (revenue/visit) – to each of these traffic sources to fully understand the potential and impact of our site improvement efforts.

As an example, one of my projects was to understand the contribution value of traffic from each of our traffic sources. The biggest hurdle to achieving this objective was the fact that we did not have any tracking to follow the user from the time it entered the experience to the time it left the experience. Depending on the size and scope of your web traffic, it could be a daunting task to collect that kind of data. However, if we tag each campaign appropriately, we can come fairly close to isolating the path. Of course the analytics software we employ will be crucial to achieving the goal (we were using Omniture and an in-house tracking solution to marry traffic and revenue data and build the user path).

Once we are able to follow the user through our site, we measured the revenue points in the path and aggregated to arrive at composite revenue/visit from a particular traffic source. Consumer centric tracking is one approach that may be useful (http://www.analyticsheaven.com/2010/04/tracking-make-it-customer-centric.html).

For instance; revenue/visit from content promotion was more dependent on user navigation and hence advertising revenue. However, search driven revenue/visit was driven more by user clicks to purchase and hence CPC revenue and/or Lead Gen revenue. We found wide variation, more than 100% between the low and high contribution values, among all our promotion vehicles. But, now we were able to employ targeting and relevant cross-sell/up-sell to grow revenue and contribution from each traffic source. Imagine the possibilities for your marketing ROI, if you were equipped with traffic source level contributions as you make decisions about where to promote, how much to spend, etc., on your campaigns. Listening to our customers and utilizing continuous feedback to update our content and products should help drive a favorable trend in traffic source contribution: http://www.analyticsheaven.com/2010/01/customer-is-always-right.html). This can be a competitive differentiator and a valuable tool in marketing portfolio planning.

Tracking - make it customer centric!

Knowing what your customers want or need, I believe, is the type of intelligence that is best obtained from internal data, rather than market studies. While the latter are a good source of market intelligence and trends, a more focused “consumer centric tracking” will ensure that we are better prepared to serve our consumers’ needs. The possibilities could be endless, from right targeting of offers, to lower acquisition costs, to greater conversion and monetization. I spoke about the customer focus (http://www.analyticsheaven.com/2010/01/customer-is-always-right.html), and now we need to ensure that our tracking is helping us achieve some of the benefits identified in that approach.

Achieving this level of granularity may be tricky, given consumer privacy issues, but we can aggregate user data in the right buckets to define our unique segments, design a scorecard to monitor appropriate metrics and create a process to quickly supplement our offerings to changing user needs.

While I talked about measuring the value of web traffic by promotion source, here the theory is that we build an understanding of user segments. For instance, if we are a content site, we may want to learn about the browsing behavior of our sports readers, segmented into say, MLB/NBA/Olympics/Soccer fans, etc. Metrics such as, time spent on the site, navigation pattern to other site sections, number of pages per visit, frequency of visits per month, shopping tendencies, awareness towards brands, etc. Following these segments may yield better returns per marketing dollars spent than a wide targeting of content and offers through the generic sports page.

What this data does is that it helps define our user experience, site navigation and product offers that are truly unique to this user segment. And we need to respect the findings to an extent that if our sports readers tell us that they are not interested in shopping (hypothetically), then we make the experience richer by showing them more content and leaving out the “useless” product ads from their pages – talk of a no-frills attached user experience. How else does one define relevance and targeting?