I have received a couple of emails asking if I have quit blogging about the topics I cover. The answer is no!
The reason I have not posted for a long while, is due to I have a new position at my work that demands a lot of time.
I will find a opening to post in the near future covering an interesting topic that have crossed my mind several times.
I just recently purchased an item using my iphone and through a website that does not “really” support handheld devices. Why not my normal computer and through a tiny screen? I was lazy and wanted to purchase the item before going to sleep.. it just felt more convenient!
Do you have an e-commerce platform that support mobile devices and do you advertise on mobile? If not.. then maybe it is time to think about it.
Watch this video from MediaMind explaining some of the awesome possibilities
I got this question today from a friend in the online business, ”how do you evaluate the performance of every online marketing channel?”
I am not going to tell you all procedures I am using to evaluate performance, because that would turn this post into a large book. However, one technique I am using is the Multi-Channel Funnels in Google Analytics. I believe it contain’s some of the most powerful conversion reports out there.
Let’s start out with one of the report’s, the Assisted Conversions report. Note! In order to use this report, you will have to tag all your campaigns using the Google Analytics UTM.
So.. what does this report show (click image above to zoom)?
Easy right?
Let’s do a quick analysis! Social-media (0.30 ratio) is a goal-scorer, that is a surprise to me let’s keep investing in this channel. TVPlay (10 in ratio) i.e. an advertising video clip e.g. Youtube / TV play. This channel is more of an assisting channel.
Now that I have introduced the macro-view over the channels, let’s dig deeper into the really cool micro part.
In the image below I have isolated a campaign during a week that cost me €3000.
(click to zoom)
Using old methods (click 2 sale) would calculate the CPO to €111.
This high CPO would make me consider to shut down this display channel (banner).
But let’s go Ninja!
The 28 assisted conversions should also be used in the calculation. But they are not ”real” conversions. Therefore, I use the 28 conversions as contribution effect to sales of 40-50% (this % contribution is set by you). €3000 / 27+(28/2) = CPO €73
Wow that is insight! Let’s continue with this campaign and give it some more time!
I hope you enjoyed this article. Please let me know if you have any questions or want share something…
What are the top 5 metrics for an online marketer? Some may perhaps believe that there would be a revolutionary magic metric that no one knows about or a secret formula to success in online marketing. I may disappoint the audience to say that metrics are of different importance depending on your business KPIs and the person observing the data. But for argument sake, here is my simple top 5 metric list with no hidden secrets:
1 Micro/Macro Conversions
- newsletter signups, transactions, customer support etc.
2 Bounce Rate
- how much do my page suck?
3 Traffic Source
- who are my best friends? always track your traffic using web analytics campaign tracking!
4 Total Visits
- if I purchase PPC traffic based on total clicks then I should also measure performance using total visits.
5 Loyalty / Visit to purchase
- helps me to evaluate campaigns, cookie time for affiliates, purchase behaviour, retarget strategies
Using these metrics with ”Advanced Segmentation” in e.g. Google Analytics, Nedstat etc. is the most important tactic for any online marketer to improve any online acquisition strategy.
No surprises with my top 5 list I guess…
But let’s skip these ”padawan” metrics from the discussion and dig deeper into this topic…let’s go Ninja!
If I would remove all the ”default” web analytics metrics (in e.g. Google Analytics, SiteCatalyst etc.) from this discussion and focus on calculated / ”integrated” metrics that you can start to use with your system! Yes YOU! I argue, that the top 5 most important metrics for online marketers are ”Real Quantity”, XCPO, Contribution CPO, Imported Profit and Campaign Profit.
1) Real Quantity 1, 2, 3 etc.
What is the point of having spent $$$ on sophisticated systems for analyzing your online marketing campaigns if it all turns out to show incorrect sales/values? The point here is that if your web analytics or adserving system will record sales using a simple conversion ”pixel/tagg” then it will always show incorrect data, due to the fact that customers will e.g. return goods purchased, regret a sales etc. The only system that can show correct sales is your backend data! Solution? Just connect your backend system with your online marketing tool (e.g. using transaction ID). Voila…your sales in e.g. web analytics will now show 1, 2, 3 correct amount of transactions!
2) Contribution CPO
Contribution CPO means that an online marketing ninja should attempt to understand how visitors interact with your different online channels before a purchase. For example, a visitor first click on a display banner, secondly an affiliate link and finally a Google PPC ad. Can your system identify the whole chain of clicks? Or only count the last click to order?..(old school!). You must understand which channels contributes to a sale and which are unnecessary to save $$$. This is possible in some adserving systems (e.g. Mediamind have a Search solution), SiteCatalyst and also in a Google Analytics release in the near future (see information here)
3) XCPO
XCPO (Exact cost per order), is the metric that will calculate real CPO e.g. if a visitor click on an Adwords link (CPC), then click on a banner (CPM) and finally click on an affiliate link (CPO) following placing an order at your site. What is the exact cost per order for this visitor? This is not as simple as explaining it because you will need proper tracking in place for your campaign, ninja analysis and measure contribution CPO.
4) Imported Profit
In systems such as Coremetrics, SiteCatalyst, Webtrends etc. you can upload profit data and connect this data to all your campaigns. Awesome! Unfortunately you can not upload data to Google Analytics but you can develop your own system to be able to do this in Excel.
5) Campaign ROI
If you have all of the above in place then you can with a simple click in your system see which of your campaigns are the most profitable. Imagine that you are running two campaigns on two different homepages and both performed 10 sales (subscriptions) each. In ”old school” analytics you claim ”both campaigns are equal”. However, using integration as described above you will be able to analyze that e.g. five of the customers from one of the campaigns upgraded their subscription to a PRO account two weeks after a purchase. Therefore, you identify that one homepage send more high quality traffic/sales and is more profitable.
Now it is your turn.. your top 5 metrics? you don’t agree?.. let me know your thoughts!
It is positive that more and more companies in e.g. Sweden (where I live), invest in web analytics tools such as Google Analytics, Adobe SiteCatalyst, IBM Coremetrics, Nedstat etc. and most important of all, invest in employing web analytics experts! While the web analytics expansion is growing, ad-serving platforms invest in their own technologies offering the same companies a different technology to track online campaigns. The issue is that ad-serving platforms collect campaign tracking data differently than web analytics platforms. While ad-serving platforms send reports to web managers based on pre-click, the web analytics platform send out reports based on post-click. This will confuse online marketers with not only the metric ”click” but also the metric “view-through”, which is not tracked (by default) in a web analytics tool.
Web analytics platforms do not track “view-through” but track campaign performance using post- “click-through” (CTR). The metric “view-through” (or VTR) is one of the metrics that many of the ad-serving platforms provide marketers with. VTR could be explained as:
a person view a banner + don’t click on a banner + purchase an item = banner will be contributed to the sale
Click-through on the other hand:
a person view a banner + click on a banner + purchase an item = banner will be contributed to the sale.
Ad-serving platforms can usually track both CTR and VTR, but web analytics only track CTR! VTR is a metric you should be careful to use in your online marketing decisions because firstly, you can never prove that a person actually saw the banner unless you ask him/her. Secondly, “the view-through metric can lead to disastrous media buying decisions…other ill effects include duplicate attribution of orders and other conversion to media that had no influence whatsoever on the final consumer behaviour” (Kevin Lee 2007). Or as Seth Godin (2010) states in a recent post, that a marketer should never “try to measure the unmeasurable media and use that to make decisions…some sophisticated marketers get good hints from their measurements, but it’s still an art, not a science”. Therefore, do not let yourself perform digital planning on gut feel based upon a single metric, but instead let VTR collaborate with your web analytics data. For example, if you are going to gain insight from a VTR campaign then it should be measured with metrics from your web analytics data. This is possible through a web analytics and adserving integration (described later in this article). I am convinced that, with the progress of web analytics expansion all online marketing decisions will eventually be based on web analytics data.
Ad-serving platforms will have to accept a situation where online marketing campaign performance will be analysed using data driven web analytics. I believe more and more web analysts in organisations will report to management the following:
- double attribution of orders
- campaign performance based upon CTR
- evaluate online budget and digital planning
- evaluate digital planning with e.g. marketing channel reports i.e. how visitors move across different online channels and from this data calculate a XCPO (Exact Cost Per Order, explained later in this article)
The analysis provided to management could cause online marketers not to afford any mistake in their online marketing spending. Or as Chris Brogan (2011) states, “If you’re not finding measurements that match your efforts and explain your sales cycle, then you’re fairly much pretending to market. Sure there are parts that you can look at like “awareness” and “sentiment,” but while those are helpful to the larger story, they don’t put money in the bank (instantly)”. This evolution could place the ad-serving platform in the shadow and online marketers would continue to listen arguments of VTRs contribution to branding and sales.
The solution for ad-serving platforms is to establish a relationship with the web analytics platform i.e. an Online Marketing Analytics solution. The goal of Online Marketing Analytics is the attempt to reach reliable online marketing data in web analytics data and optimize digital planning/campaigns that will result in an optimal online marketing spending that will increase revenue. In summary:
Exact Cost Per Order (XCPO) is calculated to understand the “real cost” for an online activity and its contribution to other campaigns in your digital media communication mix. For example, visitor clicks on a link in a newsletter, clicks on a Google ad, click on an affiliate link and then make a purchase.
Newsletter (internal resources) + Google ad (€1) + Affiliate €10 = XCPO €11 + internal costs
XCPO will enlighten the “real” cost of a campaign. The next step is to calculate campaign contribution. In this scenario (below) our web analytics/ad-serving system will track the banner to have no sales but with web analytics e.g. marketing channel report (some ad-serving systems also deliver this) you will be able to analyze the ”real flow” of the campaign!
Banner (2000 clicks & 20 orders) + Affiliate (1500 clicks & 30 orders) = 30 placed orders
First visitors click on a banner and enter a step in the purchase process that could have resulted in a sale (20 orders). But these visitors left the site but eventually clicked on an affiliate link and placed 30 orders. Out of these 30 orders, 20 may not have been placed if it was not for the awesome banner i.e. the banner contributed to the sales.
The cost for the campaign is:
Banner (€500) + Affiliate (€10 CPO = €300) = XCPO €26.66
The contribution of the banner is:
CB% =66.66% (20/30)
The CPO Contribution for the banner is:
Contribution CPO = €25 (500/20)
Formulas
XCPO = (Media Channel Cost1+Media Channel Cost2)/ Total Orders Placed
Contribution
(CB%) = Media Channel Orders / Total Orders Placed
Contribution CPO = Media Channel Cost / (CB% x total orders placed)
In the “old school” mindset all 30 sales would be attributed to the affiliate with a CPO of €10 (last click rule). In the XCPO analysis the total cost of the campaign is higher, but the marketer can see that the banner contributed with 66.66% of the orders. Moreover, the banner is distributed a Contribution CPO of €25. This will give us more insight!
If we were not measuring the contribution effect of the whole campaign then the marketer could have considered the banner campaign to be a failure. However, if we are measuring exact cost per order (XCPO), then we should also attempt to track the exact orders (XO) placed…let’s talk about technical integration…
as a client with an ad-serving platform you should attempt to establish a relationship and connect the ad-serving platform with your web analytics platforms. But don’t stop there! Create a relationship with all your backend data and your web analytics tool. It is possible to send all kind of data to web analytics platforms (if supported) such as:
Using this data the online marketer can measure the performance of his online activities with new cool metrics in web analytics and could also use the VTR data in a whole new dimension to analyse its contribution to sales. The online marketer will have ”complete” control (with the campaign tracking in place) over all his online activities because firstly, the exact amount of orders is sent from backend system and not through a code snippet. Secondly, profit data is uploaded into the platform. Thirdly, we can upload all cost data. Finally, ad-serving platforms can upload pre-click and impression data.
WIN WIN WIN
Today it is not possible to send data into Google Analytics but it I reckon it is a matter of time someone out there will build something with the Google Analytics API. In the meanwhile you can create advanced Excel documents using the same logic described.
3) Optimization Finally, in the Online Marketing Analytics model, online marketers should always attempt to optimize landing pages or banners for their online campaigns using tools such as Google Website Optimizer, Visual Website Optimizer or Adobe Test&Target etc.
During the campaign period the XCPO and contribution effects should be monitored through the sophistication of the ”new” system integration (described above) and there should always be attempts to reach a better conversion rate and maintain a healthy XCPO level!
That’s it!
I hope you enjoyed the article please leave a comment if there is something you want to add.. if you would like to know more about this process, then please contact me and I can explain more or guide you to a vendor
