eCommerce Funnel Optimization Insights with Enhance eCommerce

Enhance eCommerce empowers web analysts with a new set of reports and capabilities to gain more meaningful insights easily. if you don’t use this feature, it is high time to use it. While reading Avinsha’s blog, I found he explains the value of the enhance eCommerce setup and its possibilities. So, I thought I would share a few analyses we conducted, so you can get a better understanding and direction to optimize your eCommerce business.

After completion of enhance eCommerce, the report was as below. We can see the complete funnel

Total sessions for the selected period is 121,000, but only 37.87% visited the product pages; others may have left the website after just visiting the home page or category page. This means 62.13% didn’t visit to the product pages. All the money spent on driving them to the site was wasted with a minimal benefit. Ideally, visitors should go to product pages to understand the value of products and other value proposition including warranty and discount.  However, in isolation, we cannot say whether 37.87% is bad or good. But, as an analyst, our responsibility is to find opportunities to reduce this dropout and drive relevant traffic.

If you want to improve the funnel, you must look at the campaigns, web design and products. The below section explain the possibilities to improve the funnel along with these 3 sections.

Channel Optimization for Better Funnel

Segmentation is your best friend to understand funnel performance against each channel you are investing. You need to see multiple possible combinations to understand the segment’s top funnel conversion rates and also worst conversion rates. You can expand top converting segments while experimenting with low converting segments.

Real Case: Campaign analysis gave us the quick insight because we noted a few campaigns drive 75% visitors to the product pages from the homepage or category pages. Also, we looked at campaign with Age/Gender/Geo drill down to find good campaign opportunities.

Design Optimization

Home page & category page design layout and content should persuade users to visit the product pages. If the user leaves without navigating to the product page, it is a problem with the product or the design. To improve the funnel using design, you need to come up with a hypothesis to run CRO tests. Neuromaketing, user psychology or insight from data can be used to come up with the test.

Real Cases – We ran tests with CTAs on home page and category pages, so we were able to improve conversion rates by 20%, on average. Also, we concluded a few tests with positive results related to “Price Display”.  In addition, larger images tend to get more CTR on the category pages. By having a human image on the home page, it might increase the dropout rate because it engages with the subconscious mind and the visitor may forget their initial intention. Read more about human face psychology here –


Product Optimization

This purely depends on the product. If your product is rubbish, people don’t bother reading more about the product. When you onboard a product, you can use signals from previous sales data to forecast the attractiveness of the product.

Real Cases: We can play with the sorting order to decrease the risk with product attractiveness. We can show top selling product first, which will address the larger audience. Further, we can use machine learning algorithm to show the best possible product based on visitor attribution and behaviour.

Basically, there are unlimited possibilities where you can optimize the funnel and drive visitors through the funnel. Don’t limit to all visit to product page funnel; you can use the same methodology to optimize the entire funnel.

Tallest Christmas Tree – What’s in it for travel marketers

The tallest Christmas tree was completed successfully with a Guinness record. Every time I hear about the tallest Christmas tree as the big thing in Sri Lanka, I have the same question: “What’s in it for travel marketers?” This is a rare opportunity, where marketers can reach potential customers who are reading about this digitally via multi channels, since all the top new/media sites captured this. While we have the awareness of Sri Lanka, marketers should reach potential customers in real-time with targeted content.

If you go to Google News and search for “Tallest Christmas Tree”, you will find the all the media coverage including these top news websites:


News results for “Tallest Christmas Tree”


Not limited to the above websites, this tallest Christmas tree got coverage from almost all the countries with a positive note about harmony.

As a part of this coverage, Google Trends for “Sri Lanka” hit the highest search trend in Google within the last 5 years, and this explains the success of this effort.

Google Trends for “Sri Lanka”


This made the perfect condition for travel marketers to focus on their campaigns and offering to capitalize on this rare situation. Based on numerous campaigns and experiments, I came across a digital marketing funnel to get the maximum out of the situation. This will give you a quick guide to plan your campaign strategy, based on your maturity and objectives. But if you are an experienced marketer, this will give you a heads-up, and you can use your own creativity to launch a silver bullet.


1.    Google Display – Managed placement in top news sites


Google lets you pick the URLS or website where your ads should go. You can target all the top news websites and their posts about the Christmas tree to show your ads. Almost every news website supports Google display network. Also, you can leverage Google Display features to narrow your targeting, such as GEO, device, time, intention etc.

Also, you can personalize your ads to make them more prominent. Customize your travel package, including a visit to the Christmas tree and taking a selfie to witness the Guinness record and don’t forget to include this on your ads, which you will show in news websites.


2.    Google Search – Searchers for Tallest Christmas tree


There are plenty of searches for the Sri Lankan Christmas tree, so you can offer your customized travel package to grab their attention. Keyword planner indicates every year there are searches happening for the “tallest Christmas tree”, but unfortunately, we don’t have visibility to this year yet. But it is time to capitalize before it is too late. The below graph shows the search volumes for the tallest Christmas tree.

Search Volume for “Tallest Christmas Tree”


3.    Generic Re-marketing


If your sole purpose is to drive conversion, the abovementioned tactics have low probability, so you can stick to remarketing campaign. You can target visitors who visited your site with display remarketing and remarketing list for search ads to entice visitors to convert at your site and don’t forget to add Christmas tree to your offering.


4.    Personalized Remarketing


There are unlimited user behavior combinations you can use to target potential customers. Segment your audience based on their behavior; if someone engaged with “Yala Safari/Hotel”, you should segment them to create a specific remarketing campaign. Your ad should talk about “Yala” + “Christmas tree” and ads should appear on News websites that talk about Sri Lankan Christmas tree as mentioned above. In this manner, you can target people who have already shown an interest in traveling to Sri Lanka with very relevant ads in a relevant place.


Hope you got an idea about the potential, but you must act before it too late to capitalize on this rare opportunity. Don’t limit to the above mentioned channels; give your attention to social channels too, including Facebook, Twitter, and YouTube.

Predict Conversions using “Pages per Sessions” (Most Neglected KPI)

When we talk about engagement, pages per session is a key metric to focus and measure. Pages per session reflects how many pages visitors view during their session, which means more engagement. If a user really wants to engage with your website, he may visit more than one page. Also, this is a good indication of your offering, target and campaigns. Catherine and Amanda define customer engagement in their research as “the process of developing cognitive, affective and behavioral commitment to an active relationship with the website”


will pages per session bring money?

However, the problem with the “Pages per session” is that you need to understand how it is going to affect your bottom-line. As an example, a business that needs to drive more phone calls will focus on lower pages per session and more calls. Before you set up micro goals(Engagement goals – pages per session, time on site, button clicks, video views etc. ) you need to verify its relationship to the bottom line or macro (High level goals – Leads, transactions, orders etc.) conversion. If you have Macro goals and Google Analytics in place for a couple of months, you can see the relationship with engagement and your bottom-line. If you can prove the relationship, then you can use “pages per session” as an earlier signal to evaluate your marketing campaign effectiveness.

Prove the concept

We did a similar analysis for an e-commerce business. They have been in the industry for a long time and e-commerce conversions tracking was set up to capture transactions. However, the problem is that some products take over 60 days to convert from the first visit. So, it is difficult to wait 60 days to optimize a campaign.   The best solution for this to rely on micro conversion with the proven link to revenue or conversions. We did the analysis as below.

We considered the first visit and pages per session for the same visit also compared against the conversions rate including future visits. The following image will demonstrate the logic behind the terminology:

Linear Relationship

If pages per session during the first visit can predict the future conversion rate, we are in a good position to optimize the campaign based on pages per session. We got this from the analysis:

Y Axis – Conversion Rate, X Axis – Pages per session

We drew a scatter plot diagram to see the relationship between variables. Based on the graph, a positive relationship can be identified. The equation describes the relationship so we can come to the following conclusion.

If you increase “Pages per session” by 1, the conversion rate will go up by 0.11%. which is pretty amazing.

The conclusion is we can optimize the campaign based on engagement, so we don’t want to wait months to gather conversion data.


Real Case Study

Based on above finding, we wanted to improve the engagement of a client, so we conducted landing page analysis to see commonalities between high engaged pages vs. low engaged pages. Interestingly, we found that visitors who landed on category pages had engagement higher than the visitors who landed on product pages. This is simply because people who are visiting the site for the first time want to see more products and compare. Basically, we directed new visitors towards category pages so they could engage with the website.

Once the visitor engaged with the site, he has an emotional relationship with the website, so it is highly probable they will take action in the site or remember the brand. Hollebeek explains in his research, in 2011, customer engagement with a brand as a “mixture of intellectual and emotional interactions characterizing customers’ contextual and brand-related state of mind”.  So, we have proven evidence about the engagement and its impact on brand recall. Research conducted about “An Experimental Study of the Relationship between Online Engagement and Advertising Effectiveness” by Bobby, Edward and Ute found a positive relationship between engagement and ad effectiveness.

Invest time and effort on visitor engagement but don’t limit to “Pages per session”. You can also focus on bounce rate, time on site, site search usages and other engagement activities. Feel free to add any comments or suggestions.

Optimize Campaigns for Future Conversions – Proven Formula for Optimizing Nazis

Campaign optimization is a major task and responsibility of a campaign team because optimization logic decides the return on ad spend at the end of the month. Optimization logic should contain a lever to manage CPA/ROAS whenever necessary. It is difficult for the optimization team alone to create an optimization logic with a flexible lever to maintain CPA/ROAS. It should be a combination of ad platform features/data and analytics insights. (Read guide to get insights understanding commonalities)

Ad platform should be able to provide CPAs/ROAs for narrow segments and able to stop or enable a particular segment. Facebook provides numerous segmentation facilities to dissect your report to understand CPA/ROAS at a very granular level. It contains age/gender segmentation, other demographic and interest segments. Adwords provides keywords, placement, topic, in-market, affinity and other segmentation capabilities. Look at each segment to understand opportunities and leakages, so opportunities can be expanded while stopping the leakages.

Let’s assume the product revenue is $1000 and your marketing budget for a product sale is $100. Then, ideally, your CPA should be $100, but the problem is how many days you are going to wait before you stop your campaign/segment? You can stop it within one day, 3 days or within one week. Let’s say you stopped a campaign within one week, what if few users convert after 30 days, which means you stopped the campaign blindly?

To avoid such blind optimization, you can use time to conversion report from Google Analytics. It is a great resource to understand time to conversion from each of your paid channels. The below graph shows the “Time to Conversion” Report:



The above graph indicates that 40% converts at the same day while 60% of the users require more than one day to convert and a similar pattern applies to 3 days, 7 days and 30 days. If you stop the campaign within 1 day, you should factor 60% of the conversions will come in the future. Based on this logic, the following lever can be created to optimize the campaign. If you stop the campaign after three days, your cut-off CPA should be $100 (100/60) = $167, and it will factor the 40% of the future conversions. This is the perfect lever for you to optimize campaigns; also, you can do optimization after 3, 7 and 30 days.

Based on your overall CPA/ROAS, you can push the lever (Cut-off CPA) up or down. This gives you greater flexibility to manage your campaign in a more effective manner. This is just the start; you can factor the attribution model rather depending on the last click CPAs.

How do Pakistani E-commerce suffer from not having a CRO expert?

Ecommerce in Pakistan has been continuously growing over the last few years and shows a promising growth. With the 190 population in Pakistan, ecommerce players have a big opportunity to create billion dollar businesses.  In India, Flipkart and Amazon raised 1B+ dollar investments for their expansions.

Since Pakistan is an emerging country in terms of Digital advertisement and digital customer adoption, higher management should focus on building digital capabilities to catch up with the competition. Whoever is able to move faster with the technology can move ahead of the competition.


Fast Tempo Pakistan ecommerce

Since higher management has dealt with marketing for decades, their mindset should be adjusted according the technology advancement in digital marketing. But most of the time, we see higher management trying to mimic offline advertisement in digital marketing. They are especially falling in love with social shares and social likes. Since Pakistan is an emerging country and most customers are new to ecommerce, the success of a business can be defined in an equation as follows:


Business Growth = Adoption + Retentions


It is quite challenging to drive adoption from people who never experienced an online transaction, so you should have a strategic campaign to convert new users. Most of the time, it is difficult to acquire a customer profitably using digital time, unless it is a high-value product. As a solution for that, higher management should keep eye on the retention, and if you have a good business plan to retain the customer, then you can absorb the higher acquisition cost.


RealLife Fact: If you search for a hotel, Agoda will appear in the search result as a paid ad. Agoda is spending a lot of money acquiring new customers, because they know that if they can acquire a customer, they can retain him for a long time. In this case, individual hotels are struggling to compete with Agoda because of their strong retention.

If you need to win, become an Agoda for ecommerce. Kayamu is slowly moving towards that, as they have global expertise.

To achieve business growth in ecommerce, you need to have a wise campaign plan and a website that converts visitors to customers. Today, the object is to discuss common mistakes on the Pakistani E-commerce websites. These mistakes could occur due to lack of optimization expertise, usage of readymade templates, or the highest paid person’s opinion (Hippo). Let’s discuss two common mistakes out of 100s:

  • Automatic image sliders or carousels
  • All caps in navigation

Automatic image sliders or carousels

Hippos and creative directors are attached to sliders or carousels, so you can see most of the Pakistani ecommerce websites have a slider. All the ecommerce templates support sliders, so they can just switch it on.

A slider or carousal can take the visitor’s attention quickly and load into their working memory. Because of this, visitors may forget the initial intention and leave the website, without taking an action. Also, providing more choices to the visitor can create paradox of choices. Paradox of choice shows people leave the website if they are overloaded with choices.


All caps in navigation

Many ecommerce websites in Pakistan use all caps in their websites. Probably, all the templates come with capital letters in the navigation. Visitors tend see words in the navigation as boxes if the navigation uses all caps. Also, words in all caps increases the cognitive load of a reader. According to research conducted by Breland, text in all capital letters reduces reading speed by up to 20%. Poulton suggests, in his research, reading speed is optimal when uppercase and lowercase letters are used.


If you are using a template or a website without testing, these two suggestions can give you a way to start with website optimization. Whenever you have doubt about the suggestions, simply conduct a test to see the real result.

Agency’s Biggest Lie Revealed About Conversion Optimizer

conversion optimizer


Biggest Lie : Conversion optimizer is working well for lead gen campaign.

Conversion Optimizer (Target CPA bidding) is a cool feature, which uses machine learning to automatically target visitors who are more likely to convert. This is a good opportunity to leverage Google’s engine to drive more conversions for lower cost.

Target CPA bidding optimizes CPC, based on the probability to convert. Based on your campaign’s history of conversions, Target CPA bidding automatically finds the optimal CPC to maximize conversion.

But right now, Ad agencies blindly use this option and lose most of their budget unprofitably, not knowing the science and logic behind this. Machine learning algorithm analyze historical data and predicts the probability of conversion for each impression.

If an agency is using conversion optimizer for lead gen campaign, without using Google analytics offline tracking, I would say you are working with the wrong agency. Either you should stop the campaign or fire the agency.

The lead gen campaign’s main concern is the quality of the leads, because cheaper leads don’t guarantee the quality. Based on experience, it is highly probable that cheaper leads come with low quality. If the campaign receives low-quality leads for low cost, conversion optimizer automatically adjusts and provides low-quality leads. If your agency doesn’t link actual customers to campaign performance, your budget will be wasted, even without knowing it.

To overcome this issue, you need to use Google Analytics offline tracking feature and upload all the successful customers to GA as a Goal. Once you have successfully linked the campaign to the customer, you can use the attribution model to understand the actual performance of each campaign, ad group, keywords, ad etc. Finally, you can understand the campaign that provides low-quality leads and the campaign that provides high-quality leads and allocate your budget accordingly.

Web analysts FAIL not knowing this tactic to derive insights

The value of a web analyst has increased exponentially, since digital businesses are emerging all over the world and every business needs digital campaigns. Businesses can depend on a web analyst if he is doing the right job to add value to the business. He can add value to marketing efforts, product development, budgeting, and financial aspects.

But most of the time, we see that web analysts are stuck in a cave, not knowing how to advance their career/expertise after a certain level.  Two categories can be seen below.

Implementation Geeks

  • Know how to set-up and implement tracking for every element within the website and always thinking to capture new data points.

Web Reporters (Data Pukes)

  • Use Google Analytics only to create reports with minimal insights.


I came across two methodologies to get out of the cave and improve skills to provide more insights and recommendations. Segmentation and Commonalities are the best friends of any web analyst in any industry. Avinash (Google Evangelist) stressed these two methodologies in most of his articles to help people advance in their web analyst career.  Segmentation helps to start the web analyst’s career, while commonalities help to get out of the cave I mentioned earlier. This article will discuss commonalities, while separate articles will focus on segmentation.


Find insights by understanding commonalities

You always find good and bad with your campaigns, landing pages, or any part of digital marketing. Commonalities are the understanding of what went good and what went wrong. Web analysts should divide anything into two categories, good performers and bad performers. Then he should be able to understand commonalities or common characters of both segments. This helps to improve the good and reduce the bad. This is all in summary for commonalities. Let’s discuss real life examples to get more understanding.


Hubspot Sidekick

Hubspot’s sidekick product had an issue with customer retention, they wanted insights to improve retention. So they divided users into two groups, retained users and churned users, and looked for commonalities.  They found that if a user signs up with a company email address, they are more likely to retain for a long time. This is a heartbreaking insight for a web analyst and good opportunity to improve the bottom line for the business. Sidekick guys continuously pushed their customers to change their free email address to a company email.



Airbnb wanted to improve number of nights per customer. So they divided properties with short stays and long stays to get an insight. They found the properties with good images have longer stays and properties with low-quality images have shorter stays. Great, they have hired 20 photographers and achieved exponential growth.


ReapDigital facts:

Sinhala vs English – We ran a campaign in Sri Lanka and found that Sinhala (Most used Language in Sri Lanka) ads are attractive and drive most visitors, compared to English ads, but the problems with Sinhala ads are the conversion rates were low. So we wanted to capitalize the attractiveness by improving conversion rates. All the landing pages were ranked based on conversion rate and looked for a commonality among good and bad pages. We found that landing pages in Sinhala drove low conversion rate, even though ads are in Sinhala, while English pages drove high conversion rate. This simple insight helped us to capitalize the most attractive Sinhala ads and increase the conversion rate. The reason behind this behavior is people are comfortable to engage with Sinhala ads, but they feel annoyed typing in Sinhala in a Sinhala landing page to complete a task.


Real Estate Industry – A real estate campaign is launched to target people looking for retirement homes and luxury houses across all GEOs. We wanted to improve the conversion rate, so we sorted GEOs, based on interest (Micro metrics), to see commonalities of people looking for retirement homes and luxury homes. When the distance from Sri Lanka increases, people looking for retirement homes engaged more, and when the distance from Sri Lanka reduces, people looking for luxury home engaged. Based on this insight, we could allocate a budget, based on keywords and Geos and improve the conversion rate.


SEO using Data – Customer wants to increase SEO effort for a blog, so we got a list of blog posts and number of backlinks to them. We looked at commonalities with the pages that attracted the most natural backlinks, and we found that tutorials attracted the most backlinks and social shares. The reason behind this is people consumed tutorials and linked back from their blogs and included their answers in stackoverflow. We recommended to the customer to write more tutorials to attract more backlinks and increase the SEO rankings.

I believe that these examples are good enough to understand commonalities and advance your career.

Happy Analysis! More Insights!


“Human Image” on a website can kill conversions – but how?

Human Faces have been used in every means of advertisement for the past centuries. Without a doubt, human faces in advertisements attract our attention before anything else, and it increases when it is a well-known face, such as a celebrity.

Faces are the most important visual stimuli we perceive, and the brain can identify a known face within a fraction of a second. Research studies suggest the brain has a separate area, called Fusiform Face Area (FFA), to recognize faces above all other objects. Fusiform regions respond more strongly to faces than to letter strings and textures (Puce et al. 1996), flowers (McCarthy et al. 1997a), and other stimuli, including mixed everyday objects, houses, and hands (Kanwisher et al. 1997).

In a website/advertisement, human faces grab the attention of visitors above anything else, like in the real life, so marketers need to utilize this human behavior wisely to maximize the desired outcome.

Eye Movement for

mango eye tracking

Human Faces for Brand Recognition

If you need to improve brand recognition, the human face is the best tool to use. The human brain can identify the brand ambassador (or model) and link it to the brand, even before recognizing the logo. If a human face is used consistently in advertisement with the brand, it can transfer it to the long-term memory of your audience, which is important to capitalize every cent you spend for advertising.


Real Life Case: Kelly Felder is a famous fashion brand in Sri Lanka, and people can recognize the model before anything else. When people surf Facebook, they will come across a few dozen images, brands, and videos, so the image of the model plays a vital role in getting the attention and triggering the emotion.

Kelly Felder Facebook Image

kelly felder


Human Faces as a Friction

Sometimes, the human face on a website may act as a friction to the visitor, as they scan faces before anything else. Because of this behavior, visitors’ cognitive load (the total amount of mental effort being used in the working memory) may increase, and they leave the website, without completing the desired outcome. User attention is a precious resource and should be allocated accordingly.

Hubspot ran a test with human face to improve their lead conversion rate but it didn’t. Probably it occurred due to the increase of cognitive load.

hubspot ab test


Human Faces as a Slider

Sometimes, human faces on a website may act positively and increase conversion rate. We had the same experience with our tech support brand.

24/7 Techies Website

247 techies

ReapDigital Fact: We did several tests with our tech support brand website to understand the influence of a human image for our customers. Multiple tests were conducted with male, female, young, and old human images, but all the variation won against a landing page, which had no human image. Based on multiple tests and data analysis effort, we could come to the following conclusion.

In the tech support industry, people fear being scammed. So, their conscious mind (conscious mind is your thinking mind or the logical part) actively participates in rationalizing the services. If the conscious mind is unable to rationalize our service as a legitimate business, they leave the website without taking further action. The conscious mind requires a lot more effort and evidence, so it can be easily overloaded. By having a human face on the website, initial attention goes to the human face, reducing the effectiveness of the conscious mind, while the subconscious mind (Participates in emotional decisions) comes into play. Also, human faces can have a close relationship with visitors’ subconscious mind and motivate him to take the decision, based on emotions, which is immediately to get support to fix their computer.

Bart Schutz (Psychologist) mentioned that “wish list” on an eCommerce site may not have a real value to the customer, but it may reduce the participation of the conscious mind thus increase the conversion rates.

@Jtrondeau presented the below AB test case study. Variation with the social share buttons won with a 5.7% improvement. Someone can think that large social buttons may act as a friction, but in this case, those buttons may reduce the participation of the conscious mind.

AB test

There is no correct or wrong answer for the use of human faces in a website, so you need to test and understand the customer’s mind and act accordingly.

Hope you enjoy the article and all the comments and criticisms are welcome 🙂



Marketing to the “Long Term Memory” – How to Not Waste Advertising Money

long term memory marketing 

Our brain contains two types of memories, which are short-term memory (or working memory) and long-term memory. Short-term memory is the short time we remember something before either dismissing it or transferring it to long-term memory. As marketers, we don’t want only to talk to short-term memory and waste our effort. Before everything, it is important to understand the memory type and its behavior.

Short term memory can be trained using “Maintenance rehearsal” which is a memory rehearsal that stores information into the short-term memory, but it is not an effective way store information into long-term memory. Long-term memory can be trained using Elaborative rehearsal, which is an effective way to transfer information into the long-term memory. Elaborative rehearsal involves thinking about the meaning of the information and connecting it to other information already stored in the memory. It goes much deeper than maintenance rehearsal.

A few important characteristics of long-term memory help marketers to pass their message to potential customers’ long-term memory

  • Related information and Storytelling – Elaborative rehearsal involves thinking about the meaning of the information and connecting it to other information already stored in the memory.
  • Visual long-term memory has a massive storage capacity for object details. (Standing, 1973)
  • Multi-channel and multi-device – Engaging with the material in more than one way encourage elaborative rehearsal.

Related information and Storytelling

Google has a very data-led culture. However, we care just as much about the storytelling and the brand, and how we tell the world about our mission. (Lorraine Twohill, Google’s senior vice president of global marketing). This indicates that Google understands the value of passing information to the long-term memory.

Emirates airlines’ cabin crew performed during a break in the game at Germany’s Volksparkstadion football stadium in Bahrenfeld, Hamburg, Germany to create the awareness of flight safety in front of 60,000 people. This is a very effective way to communicate the brand to the long-term memory with something already in the mind, which is football.


Elaborative rehearsal

Watch the video here: Link

 Visuals – “A picture is worth a thousand words”

Several campaigns proved that image ads are more effective than text ads in most cases. However, use images wisely. An image relevant to something in a mind could be a powerful way to communicate a message, rather than a generic image.

ReapDigital Facts: Personalized image ads based on designations/company provided above average return for the invested money. Personalization Case Study Here

Among the images, the human brain has a separate section to identify the faces above all other objects, so the face is a very powerful instrument to use in marketing activities. (Let’s discuss this in a separate article)

Multi-device and multi-channel campaigns

Marketers get access to customers’ long-term memory by reaching them using multi-channel and multi-devices. This is not a new concept, but doing it right is not common. A proper nurturing process can communicate the message to the long-term memory.

ReapDigital Facts: The team found that 90% of the conversion through (Thanks to Google Analytics Cross Device Tracking)  mobile campaigns happens using the desktop. People click the ads on mobiles but complete the transaction on a desktop. (Found across few clients.)

Don’t waste your money with short-term memory, hit the long-term memory with every cent you spend.



Multi-Property Attribution Model –Hotel Marketers Can’t Ignore It Anymore

Measuring campaign success across multiple channels is essential to gaining insight to invest on each channel. Nowadays, common buzz word is multi-channel attribution, even though marketers don’t practice it. Multi-channel attribution discusses profitability of each channel while considering the channel position in the customer’s journey. Channel positions vary from the first click, last click through to view through. Finding the right attribution model is quite challenging but once you have found it, doors will be opened to maximize the revenue while at the same time reducing the cost per acquisition. Apart from the multi-channel attribution, cross device attribution provides insight to allocate budgets across devices and to optimize campaigns to generate better returns.

Now it is time to introduce a new concept to measure the profitability across websites/properties. Modern customers go through multiple properties before they eventually make a purchase. Right now it is not possible to use multi-channel attribution modeling to solve this problem, as customers don’t click on a link to move from website to website. Let’s see an example in the travel vertical below.


The customer starts the journey with TripAdvisor and reads a review then goes to the hotel website to get the information. Finally, they close the deal with as he is a loyal customer. The conversion will be attributed to the last engaged website which is and you can’t see the impact of various websites beyond with the existing analytics tools. In this case, the hotel website supports in creating the awareness (assisting) while Agoda closes the sale (last click). Multi-property optimization allows users to identify the real contribution of each websites.

It is important to quantify the impact of each customer engagement so that marketers get the opportunity to allocate a budget to each campaign. Let’s consider the same customer journey with hypothetical numbers. Let’s assume a hotel room is priced at $100 and the hotel owner pays 15% commission to

Scenario 01: Without Driving Traffic to the hotel website

The Hotel Owner makes money entirely from with no investment on the website.

 Scenario 02: Customer Drive Traffic to the Hotel website

Hotel owners expect to keep the cost per acquisition below $30 with paid campaigns. However, hotel owners don’t have a way to figure the multi-property attribution with the existing tools. The owner might’ve spent $1,000 on paid campaigns and only received 10 conversions from the website. Based on the campaign statistics, the cost per acquisition (CPA) for the campaign is $100 which is not profitable, as he previously expected. A marketing consultant would recommend him to stop the campaign immediately.


 Quantifying the Multi-Website Attribution

If we have a mechanism to quantify the multi-website attribution then it allows us to identify the true impact of the campaign. Let’s consider the impact of the above campaign. drove an additional 30 conversions. Now, we have 40 conversions as a result of the paid campaign and it reduces the CPA to just $25. Yes, now the hotel owner has a profitable campaign and the hotel owner can expand it to maximize his revenue. This is the insight that we can extract from a multi-property attribution model.


Multi-channel and cross device attribution can be analyzed through Google Analytics and other similar analytics tools in the market. However, it is challenging to build a tool to find the attribution across websites for which we don’t have access. Good tagging framework won’t help us to solve this problem unless we have a mechanism to share a unique identifier across the customer. As a solution, we can use mathematical models such as regressions analysis to figure the revenue relationship between the hotel website and the (as in the example). Apart from that, there are websites that provide the complete user journey and they are able extract user data from their add-on users. Finally a multi-website attribution is a model that marketers can no longer ignore.