Trust and transparency issues, downward cost pressures, and the increasing role of performance marketing expertise as a core competency, has led a growing number of eCommerce brands to bring advertising under their own roof. 

Agencies can be expensive. Their costs have to price in hours worked on proposals that don’t convert, training for their teams, and non-billable overheads such as office space and support functions. Depending on your geography, bringing in-house could be more economical, and provide you with dedicated resource when building your performance marketing program.

Look to in-house if paid media is within the top two or three transaction drivers. However, don’t fall for the misconception that hiring one or two people to take care of a strategically important function is “in-housing”. It actually requires a broader investment, not only in people and development, but also investing in your tech to truly make a difference, and match performance of a specialist top-tier agency. 

In-housing to develop unique capabilities is very much worth it, but if the primary motivation is purely to save costs, then you may quickly find that stagnating performance due to under-investment is a very real risk.

Whether you’re a startup eCommerce brand or an established one looking to scale into new markets, understanding your target audience is just about the most important thing you need to think about during the campaign planning process. Relevancy is everything in marketing, and by being laser targeted on who you’re talking to, you’ll impact every possible metric – in particular engagement and conversion metrics.

An eCommerce customer persona – also known as a buyer persona – is essentially a document that contains fictional avatars of your ideal customers. They describe your customers’ needs, preferences, motivations, amongst other things and can be used to inform ad targeting, content marketing, and even product development.

Personas are an illustration of your ideal customer. They’re a useful starting point for coming up with ideas around targeting, ad creative etc. as they provide an overview of your target customer. For example, if you’re working with several agencies across different channel workstreams, having a centralised view of who you’re actually looking to target can ensure greater alignment and integration across your performance marketing, email marketing, SEO, and website content.

Once you have crafted a detailed customer persona, you can use ad platforms like Google and Facebook to target them based on demographics and interests. For example, we can target people that like interior design as an affinity audience in Google, based on things like their search history and the websites they visit most often.

You might have also heard of audiences within the scope of digital marketing. Personas and audiences are interlinked. You create the persona first, which then feeds into how you would target them – audiences – across ad platforms.

Regardless of the type of business you have, its size or scope, you need to create at least three customer personas to guide your marketing efforts. Sure, you can skip this step, but your targeting will be overly broad, your messaging will be loose and the website traffic that you do manage to generate won’t convert as efficiently.

NEXT STEPS

• Read this article on developing customer personas
• Checkout the blimpp Customer Persona Tool

Whereas keyword targeting will always be a core part of paid search, layering audiences allows for greater ad personalisation and insights into user behaviour.

There are several audience types that can be applied to search campaigns, in order to improve targeting and overall relevancy:

In-Market Audiences
In-market audiences consist of Google users who are ‘in-market’ for specific products and services, and therefore may be more likely to to buy.

Affinity Audiences
Affinity audiences are slightly broader than in-market, as it reaches users based on lifestyles, passions, and habits. With over 80 unique personas based on lifestyle and interests, affinity audiences mimic the depth and breadth of TV-style audiences, so marketers can engage with precise audiences at scale.

Detailed Demographics
Google has long allowed advertisers to view and target users by demographics with bid modifiers based on age, gender, and household income. In the past couple of years, Google has gone further by introducing detailed demographics. They are Parental Status, Marital Status, Education, and Homeownership Status.

Similar Audiences
Facebook has lookalikes, Google has similar audiences. For Google, this is where it targets users who have similar search behaviour to users in your remarketing lists. Google is vague as to what determines ‘similar search behaviour’, but if you are looking for new customers, and already have extensive remarketing lists set up, then this type of audience is certainly worthwhile testing.

Remarketing Lists
When it comes to convincing the people who visit your site to take action and convert, one visit is typically not enough. Remarketing gives you a chance to bring back indecisive customers with tailored ads and offers that relate to the experience they had on your site. You can create remarketing lists based on website visitors, YouTube users, or customer data in the shape of CRM data.

Combined Audiences for Search
Previously, audience lists were pretty much standalone, with limited ability to combine or narrow targeting in order to achieve maximum relevancy. Then, in October 2019, Google launched Combined Audiences for Search, allowing advertisers to create various audience combinations, blending a range of sources including affinity, in-market, remarketing, similar audiences, and detailed demographics.

Given the emphasis on algorithmic bidding and machine learning, accurate conversion tracking is must-have for eCommerce advertisers looking to efficiently scale their Google Ads activity. There are two main ways to track sales within Google Ads: the dedicated Google Ads conversion tag, and importing transactions and other goals from Google Analytics.

Both approaches have their merits and drawbacks, and you’ll need to look at these closely when deciding on the most suitable method.

Tagging
Google Analytics only requires a single tag on all the pages of your website in order to track people engaging with your content. While the Google Ads conversion tracking tag must be placed on the individual page that people view when a conversion has taken place. This means if you have multiple conversions, you’ll need to configure multiple Google Ads conversion tracking tags to collect data.

Tracking Conversion Value
Both Google Ads and Google Analytics allow you to assign a static dollar value when configuring a conversion. However, Google Ads also gives you the option of tracking a dynamic dollar value for each of your conversions. This lets you pass the exact value of the conversion through to Google Ads. If you want to track a dynamic value with Google Analytics, then you will need to implement eCommerce tracking instead of using a goal conversion.

Attribution
If you choose to use Google Ads conversion tracking, you’ll only be able to report on the performance of your Google Ads campaigns and you won’t see the performance for your other marketing channels. So even if you use the dedicated Google Ads conversion tracking, you’ll also need to configure the conversion inside Google Analytics in order to understand how each of your marketing channel performs in relation to one another.

View-Through & Cross-Device Tracking
When you import conversions from Google Analytics, you won’t be able to make use of the ‘view-through conversion’ or ‘cross-device conversion’ metrics inside Google Ads. These metrics are only available directly inside your Google Ads account if you’re using Google Ads conversion tracking. So if you add these columns inside your account, you will see zeros reported for your display and YouTube campaigns. Instead, you’ll need to use the attribution reports inside Google Analytics.

With the above in mind, you may be better off incorporating Google Ads conversion tracking if you:

• You extensively leverage Smart Bidding, and want to provide the Google Ads algorithm with as much relevant data as possible for it to optimise your bid strategies.
• You want to run campaigns across display and YouTube, would like to accurately report on the view-through performance of these campaigns from within Google Ads.

Conversely, Google Analytics may be the better fit, if:

• You want a quick and easy way of implementing conversion tracking, with minimal need for deploying additional tags.
• You already have eCommerce tracking configured.
• You want to gauge how effective your Google Ads performance is vs. other channels

NEXT STEPS

• Decide on which conversion tracking is most suitable to meet your needs.
• If you want to move forward with Google Ads conversion tracking, check out this implementation guide.
• If you want to import Google Analytics conversion tracking, see this guide.

Optimising towards the mobile user is now critical to fuelling eCommerce growth. Site speed, responsive design, mobile image functionality are just a few of the key factors to consider.
 
Mobile Page Speed
Page speed measures the time it takes the content on a URL to load. Site speed, on the other hand, represents how your site is performing overall.A time delay can affect sales, too. Through client experiments, Crazyegg discovered that speeding up a page by just one second boosted conversions by 7%. That means if a site makes $100,000 per day, a one-second improvement generates another $7,000 daily, and the opposite is true too.
Slow-loading sites also negatively impact your traffic generation. Within paid search, for example, slow landing pages lower your Google Ad Quality Score, resulting in higher Cost Per Click. SEO experts will also tell you that slow pages hurt organic rankings across all devices.
There are a few tried and tested ways to speed up your mobile page speed performance:

• Use fast and reliable hosting
• Use a content delivery network, especially if your traffic is spread across multiple countries
• Remove redundant tags fro your website
• Prioritise testing and optimisation of your mobile performance
• Decrease thumbnail image sizes
• Ease up on homepage hero slides
• Be selective of which apps you use on your site, given that this has a direct impact on page speed
• Minify your code

Image Functionality
People are familiar with gestures like double tapping and pinching in order to zoom in to images on mobile. During a mobile eCommerce usability study, the Baymard Institute found that, quite naturally, shoppers want to be able to inspect the product thoroughly and are concerned about the small details. Users who were interested in purchasing a product but couldn’t explore it by zooming in didn’t feel comfortable buying and often didn’t commit to the sale.
 
Image Resolution
Low-resolution images are the equivalent of no-zoom images. Users are interested in seeing the details of a product clearly and need high-resolution images to do so. The option to zoom in on a low-quality image is pretty much useless, as it doesn’t allow the user to see details.
 
Product Bookmarking
The small screen size of mobile devices is a primary roadblock when it comes to mobile eCommerce. The Baymard Institute observed that more than 61% of mobile users will often use their desktops to complete a purchase rather than their phone. A “save shopping cart” feature reduces the number of cart abandonments and enables shoppers to save items for a later purchase. Persistent shopping carts allow customers to continue their shopping without the need to search for the desired product on their return.
 
Suggestive Search
Address lookup and validation accelerate the checkout process – various APIs, such as Google Places, enable easy implementation of this feature.
 
Mobile Search
Keeping the search bar easy to access and above the fold gives shoppers a way to search for products effortlessly. Users typically search the application at the beginning of their shopping journey, and there can be instances when displaying a search form later in the journey may harm the UX. For example, during the checkout process, search is no longer useful and will only distract because it could make users lose focus and forego their commitment to the purchase.
 
Provide Smooth, Easy, and Speedy Checkout
Studies have shown that the checkout process is the most stressful part of completing a purchase online. The Baymard Institute reports a rate of 35% for cart abandonment because of the requirement to create an account before purchasing. In order to decrease cart abandonment rates, it’s important to design an easy, fast checkout process that doesn’t ask shoppers to register first. To promote user engagement, label each step of the checkout process by numbering and displaying the currently active step. Use visual representations like progress bars to display advancing forward in the checkout process.
 
NEXT STEPS

• Check you Google PageSpeed score.
• Use the Technology > Browser report in Google Analytics to identify the most relevant mobile browsers.
• Use user testing to highlight conversion bottlenecks across mobile.
• Fix your images, with particular attention to quality, detail, and size.
• Use a tag management solution, like Google Tag Manager, to clean up your tagging across your website.

Selecting the right keywords remains a critical task when refining Google Ads campaign performance.

Keywords are words or phrases which trigger against search queries relevant to your business. The keywords that you choose are used to show your ads to people, hence why this is such an important element of building a robust Google Ads program. Select high-quality, relevant keywords for your campaigns to help you reach only the most relevant people, who are more likely to buy your products.

When it comes to thinking about keyword lists, it helps to think about the goals of your Google Ads campaigns. If you’re a DTC brand that’s looking to leverage Google Ads for primarily defensive reasons, such as to appear ahead of competitors bidding on your brand keywords, you may choose to focus primarily on brand and a handful of relevant category-specific terms. The risk with a predominately defensive strategy is that you would only generate a handful of clicks on your ads, and the prospects of scaling your Google Ads activity would be fairly limited.

A more expansive approach would be to bid more broadly on category search queries, across a multitude of ad formats. For example, you may look at Dynamic Search Ads or Google Shopping as a means to expand reach, and drive as much visibility as possible within your vertical. In doing so, you’re likely to facilitate more opportunities to drive relevant traffic, however be careful, as going overly broad may drive website visitors that are also less probable to buy.

Achieving the right balance between traffic and relevancy is key, so give yourself the best chance of profitable growth by choosing the right keywords across the right target audiences.

NEXT STEPS

• Review your search terms report for keyword expansion opportunities.
• Implement Dynamic Search Ads to help drive incremental traffic opportunities.
• Broaden product-level visibility, with Google Smart Shopping.

Google Ads Smart Bidding uses machine learning to optimise bidding within each and every auction.
Using data signals like time, user device, location, and ad creative performance, Smart Bidding can figure out the likelihood of conversion and then optimise bidding around those data inputs. This is not only a much more efficient way of managing bids, it also frees up countless hours which can be reinvested elsewhere, on things like strategy, testing, and innovation.
There are several different bid strategies you can use depending on what your goals are.
Maximize Clicks
Maximize clicks is a Smart Bidding strategy that be used to increase the number of clicks your ads receive. This strategy works best across brand campaigns, where the main goal is to drive traffic. the main benefit of this strategy is that it can lead to lower click costs, which can be useful if you’re seeing extensive competition across brand search landscapes.
Maximize Conversions
Maximize Conversions is a Smart Bidding strategy that optimises bidding to get the most conversions per campaign. Using auction-time insights, Maximize Conversions uses conversion-centred data signals to drive the most conversions for your budget. This is a useful strategy if you find that many of your campaigns are limited by budget.
Target ROAS
Target ROAS works similar to Mazimize Conversions, however the key difference is that it optimises bidding towards a specific Return On Ad Spend target. Using the data within your conversion values column, Smart Bidding is able to predict future values through conversion tracking. Google recommends having a history of conversions before using Target ROAS, with most campaigns needing at least 20 conversions in the past 45 days to be suitable for this particular strategy. It’s also important not to set unrealistic ROAS targets, using your historical account average as a starting point, and working up in 20% increments from there. Bear in mind that Target ROAS is likely to limit your traffic volume, as it will favour certain auctions over others.
Target CPA
Similar to Target ROAS, Target CPA seeks to drive as many conversions as possible within a specific constraint, in this case based on a Cost Per Acquisition target. You’ll need to drive past conversion volume within your campaign for this strategy to be effective, and it’s highly likely that your traffic volume will decrease, as the algorithm learns which auctions are likely to drive conversions. For most eCommerce advertisers, its recommended that you choose Target ROAS over Target CPA, unless you have uniform pricing across your products.
Target Impression Share
Target Impression Share, as the name suggests, is different from the other bid strategies as it focuses on optimising towards Impression Share above clicks and conversions. This strategy is designed to help brand advertisers who are looking to drive as much visibility as possible, within Google search auctions. Target Impression Share works best across brand campaigns, where you want to appear as much as possible whenever someone searches fro your brand. There are three search page locations you can choose to optimise towards – the absolute top of the page, the top of the page, and anywhere on the first page.
When setting up a Smart Bidding strategy, expect a learning phase of anything up to a couple of weeks, depending on the traffic and conversion volume of your account. If you have relatively modest number of clicks of conversions, you may see a longer learning phase, as the bidding algorithm figures out what works and what doesn’t for your business. On the other hand, you may experience a reduced learning phase if you have a higher volume account, and are providing Google with lots of data to base learning on.
NEXT STEPS

• Decide on which bid strategy is best for you, based on your specific goals within each campaign.
• Implement your desired bid strategy, allowing as much time as necessary for the algorithm to learn your account trends.
• Optimise you bid strategy over time. It’s good to work in 15% to 20% increments when it comes to the target goal of your campaign. Depending on the amount of traffic your account generates, allow a week or so between changes, so that you are giving the algorithm enough time to adapt to the new target.

An attribution model is the rule, or set of rules, that determines how credit for sales and conversions is assigned to touchpoints in conversion paths. For example, the Last-Click model in Google Analytics assigns 100% credit to the final touchpoint that immediately precede sales or conversions. In contrast, the First-Click model assigns 100% credit to touchpoints that initiate conversion paths.

You can use the Multi-Channel Funnels Model Comparison Tool to compare how different attribution models impact the valuation of your marketing channels. In the tool, transactions and revenue for each of your marketing channels will vary according to the attribution model used. A channel that predominantly initiates conversion paths will deliver higher revenue according to the First-Click attribution model than it would according to the Last Interaction attribution model.

If you’re still using Last-Click to report on channel or campaign performance, you’re most likely under-valuing those channels and campaigns which drive proportionately higher First-Click performance – e.g. Display, Google Shopping, Facebook Ads. Pausing these campaigns, or lowering their budgets, could have a potentially adverse impact on your entire acquisition funnel. Last-Click attribution robs you of essential data about what’s working at the top-of-funnel.

Changing your attribution model can feel like a shot in the dark. You’re not sure what will change and the stakes are high. The impact of your attribution model effects everything from how you report revenue into the business, to how you optimise and allocate ad spend and resource between channels.

NEXT STEPS

• In order to make it easier for you to predict what will happen when you change your attribution model, check out the Google Attribution Model Comparison Tool.

Facebook dynamic ads automatically show the right products to people who have expressed interest on your website, in your app or elsewhere on the Internet. Simply upload your product catalogue and set up your campaign once, and it will continue working for you for as long as you want – finding the correct people for each product, and always using up-to-date pricing and availability.

Unlike other Facebook ad units, Dynamic Product Ads uses an ad template that automatically uses images and details from a product catalogue to advertise. No more manually creating thousands of ad variants for different products, as Dynamic Product Ads removes much of the heavy lifting.

To get up and running with Facebook Dynamic Product Ads, you’ll need to do the following.

Install The Facebook Pixel
The Facebook Pixel allows you to create highly relevant audiences based off pixel events, for example, people that have viewed a product but not purchased. It also provides the basis of lookalikes, which will be a key ingredient if you’re planning on using Facebook Dynamic Product Ads for prospecting, and not just retargeting.

Create Your Product Catalogue
This is essentially the link between your eCommerce website and Facebook, and provides a structured view of your product data, much like product data feeds for Google Shopping.

Segment Products Into Product Sets
A Product Set is essentially a group of products that share similar characteristics, such as a product category, brand, or a particular price point. You can get very clever with how you use your Product Sets to advertise particular products to particular audiences, but for general retargeting, you’ll only need a single Product Set. Where it gets more complicated, is when you want to advertise carefully curated products based on up-sells and cross-sells.

Choose Your Audience
Generally, Dynamic Product Ads work best if your going broad – think lookalikes of your best customers – or super-narrow, such as cart abandoners, product viewers, and product purchasers.

Create Your Ad Template
Now to the easy part, given that most of your product information is populated from the product catalogue. You can also use “dynamic placeholders” in the ad text, which may include things like product name, price, description, and landing page.

NEXT STEPS

• Connect your products to a third-party app, such as GoDataFeed, to automatically create a Product Catalogue from your product inventory.
• Download our 44-point checklist on Facebook Dynamic Product Ads.  

Facebook value-based Lookalike Audiences can help you to reach more people that resemble your high-value customers. By telling Facebook how much your customers are worth to you  in terms of the Lifetime Value, their smart algorithms can help you find more people similar to your customer base.
The simplest way of creating value-based lookalikes is to upload a spreadsheet with high-value customers, however, this is by no means the only way to optimise towards LTV. You can also create lookalike audiences based on pixel events, which has some compelling advantages over a simple file upload:

• They separate out your best and worst customers by value. For example, if you’re a fashion brand you might want to create audiences based on high ticket formal dress purchases, and not lower value purchases, such as small accessories.
• The recency of the data is also taken into account. Data from the last 60 or so days is analysed with more weighting given to recent transactions.
• The frequency of purchases is another factor taken into consideration.
• Previously we could create a source audience based on value OR recency OR frequency but never before have we been able to combine all three so easily.

NEXT STEPS

• Upload a Custom Audience data file with a column for LTV, or set up your pixel or app to capture events that contain value, such as purchases.