Let’s take a first look at the branch of Marketing Analytics known as Digital Marketing Metrics (or Web Metrics), which sit within the category of Sales Metrics (see earlier posts here and here). That is, they relate specifically to sales, and their measurement and improvement leads to increased sales revenue for the business. They are also a kind of Dynamic Metric: one which is likely to change significantly in a short timeframe (within one month or less), and one which can be significantly impacted in the short term by actions the company can take now. For example, making small tweaks to the customer journey through the website may increase additional purchases, or shortening the number of steps, and hence time to checkout reducing cart abandonments, and so on.
Sales metrics in general are used to review how well the business is identifying potential customers, engaging them or capturing their attention, communicating to them about its products or services and marketing offers, and convincing them to buy its products or services. Web metrics do the same job, but specifically in the context of marketing and selling online which has its own unique aspects (such as monitoring online advertising conversions, tracking visitors’ clickstreams, optimising web pages to minimise friction and drive sales).
Daniel Egger [1] defines marketing as a “deliberate, measurable process” used to create and increase the target audience’s “awareness and interest” in the company’s brand or offering, where that target audience shares in common a specific set of characteristics (usually demographics such as age, gender, geographical location, etc.).
Marketing metrics help the company target more closely those prospects and potential customers who are most likely to respond to the company’s offering. thus spending its limited marketing spend most profitably and effectively. Such metrics help the company to hone in on the perfect target; the person most likely to engage with the company and its offers, buy its products or services, and become a loyal long-term customer who returns to buy again and again.
Online marketing:
In online marketing this honing in on a target demographic or target market is fundamental to success (unless you have an unlimited marketing budget at your disposal), and is often referred to as ‘niche marketing’. The trick is not to compete against all the other major players and try to get access to every potential customer in a broad market area (e.g. all pet owners in the world) but instead to focus in on a small sub-section (or niche) that you will be well-placed to serve really well (e.g. providing advice and guidance to huskie owners in the UK).
To do this it is necessary to identify (1) what is the competitive differentiation or specific uniqueness of your product or service, and (2) who are your ideal target market and what specific set of characteristics they will share: they love the product/service in general (pet-lovers), they have need or preference for your specific version of the product/service (thinking about getting, or already have, a husky as a pet, and live in the UK), there is no other product/service provider available to them locally, and they can afford your price point.
If you work with the already-available demographic categories (age, location, household income, interests, etc.) you will be able to hone in on the specific characteristics of your ideal target market. This can be done through traditional direct marketing techniques of targeting ads to relevant magazines or newspapers, buying lists of target names and sending out a mail piece, etc..
The process is made so much easier online though, through the advent of social media such as Facebook (find groups of Husky-lovers and geo-target ads to its UK-based members), Twitter (find people talking about huskies) or YouTube (find those sharing and commenting on husky videos), or with targeted advertising through Google AdWords.
Google AdWords:
(Note: other online search advertising is available.) Google Adwords allow companies to target online advertising to a specific target audience (people searching for a particular subject on Google). They do this by bidding in an auction-based pricing system for access to specific AdWord terms or phrases which can be aligned with the subjects/phrases customers are searching for within Google. This is known in the industry as PPC (‘Pay per Click’).
Pricing varies for search terms/ad keywords, based on commerciality and level of competition for a specific keyword. AdWord price is denoted as CPC (‘Cost per Click(through)’). Companies will bid up to a their own defined ‘Max CPC’ for any given keyword they wish to use. The actual price they end up paying if successful in winning the auction is the ‘Actual CPC’. Actual CPC normally lies in the $1-$2 range, but heavily contested topics such as insurance or mortgages can have CPC’s hitting the $45-$55 price range.
This seems cost-effective in light of the company’s probable internal cost to recruit a new customer (which we would usually expect to be far higher than a few dollars), but the cost of the ad click is due to Google regardless of whether the traffic clicking through on your ad even bothered to read the sales page let alone convert to a sale. So optimising for click-through’s alone will not be enough; one also needs to optimise for conversion on the landing page or sales page.
In the meantime, Google is busy optimising it’s own revenue from showing AdWords on its search pages, and will always rank more highly that advert which is likely to earn it most ad revenue, based on the Actual CPC for the ad x some quality score for the ad. The ad’s quality score will be calculated based on some internal Google algorithm to account for factors like: how likely customers are to click on the ad, how relevant the ad is to the search term the customer searched for, and the landing page experience the customer is likely to receive upon clicking through (bearing in mind the information they originally searched for). It’s possible for Google to check these factors by parsing (effectively, reading) the text of both the ad and the resultant landing page, and ensuring they match with one another and with the keyword (as related to the customer’s search term).
What this means in practice is that AdWord adverts and landing pages should be constructed as far as possible in natural language, similar in subject matter, quality and relevancy to those websites and webpages being organically ranked most highly by Google for that given search term and associated ad keyword. So keep your landing page natural, relevant and high quality, with substantive information on our product or service, which is also directly and clearly related to the subject of the original search term. By doing so we should achieve a high quality score for our ad and hence have it appear high up on the list of sponsored ads on the Google search page.
AdWord metrics:
The kinds of metrics likely to be tracked are – for all those prospects who visit our landing page through a specific AdWord ad:
- the Actual CPC (how much it cost to get a prospect to view our landing page);
- how many visitors in total clicked through the ad to visit our landing page;
- how many actually converted to take the required action on the landing page (e.g. subscribe to our newsletter, download a free e-book in exchange for a name and email address, purchase a specific product or service);
- the % conversion rate, that is how many bought (3.) as a percentage of the total who clicked through (2.)
- the acquisition cost for each new customer for that particular advert, which is calculated by dividing the Actual CPC (1.) by the % conversion rate (4.). An alternative calculation is: the total cost of all click-through’s for the advert (1. multiplied by 2.) divided by the number who converted (3.).
Both ways of calculating 5. give the same answer, that is, the total spend for each new customer acquired (or converted) with that particular advert. This metric can provide a relative ‘effectiveness score’ for that advert’s ability to convert, as compared to other advert which might be run at the same rime for the same landing page, if we set up some kind of A/B test to randomly test the two different AdWord adverts alongside one another.
Lifetime value metrics:
Once we being to build up a larger volume of data, we can begin to get a sense of how many customers (on average) will only buy from us the one time and never be seen again, compared with those customers who come back again and again (repeat customers). First we’ll need to have a large enough sample size of conversions on our landing page.
From here we can start to look at the average pattern of purchases (or conversions) over time, by looking in detail for each customer line by line what their purchases were and when. We can use this to generate a Life Time Value (LTV) metric, the average life time spend per customer (applicable across the whole population) expressed in terms of the “present value of all future revenues generated from that customer”.
There are a variety of different ways LTV can be calculated dependent on business context (some methods shown below, or see KissMetrics for a useful infographic on the subject).
(EXAMPLE FORMULAS HERE)
Max CPC:
If we know what the life time value is likely to be for a given demographic/target audience, then we can calculate what should be the maximum we are willing to bid and pay for an AdWord, since the cost to acquire a new customer should never be greater than the expected lifetime value of that customer if we wish the ad campaign to be profitable.
Example:
LTV = £300
Conversion Rate = 2.9%
With MaxCPC / Conv Rate < LTV
Then MaxCPC < LTV x Conv Rate
MaxCPC < £300 x 2.9%
MaxCPC < £8.70
So we can pay up to £8.70 for the AdWord and expect the ad campaign to profitable, or if we pay exactly £8.70 then the campaign will be expected to break-even.
In reality, since our LTV value is likely to be an average (and, dependent on the sample size and value of the data used to calculate it, there may be some error contained within meaning our actual LTV may come in lower than the inferred/average £300. In this case it would be prudent to set our MaxCPC lower than this theoretical MaxCPC of £8.70, perhaps setting it (say) 10% below this level to provide some headroom in case of statistical error, giving a revised MaxCPC of £7.83.
The above calculation shows what our maximum spend on ad campaigns should be, based on the expected total revenues from any newly acquired customer, over the whole lifetime of them buying from the company. The danger here is for the company to get carried away and spend an unsustainable (and unsupportable) amounts of money upfront on advertising and new customer acquisition, long before all that lifetime revenue crystallises as sales and, ultimately, cash receipts.
A more sustainable approach will be to base the calculation for MaxCPC on the likely spend per customer in their first year of trading with the company (i.e. “present value of the first year’s revenues generated from that customer”). In this case the prudent goal is to minimise cost of new customer acquisition – while at the same time maximising first year sales revenue – such that the company is able to generate a profit from that customer within their first year of trading. Where the cost to acquire is higher than the expected first year revenues, there is likely to be a cash flow risk (as well as being unsustainable over the medium to long term).
Customer segmentation:
All the above metrics can of course be modified to deal with more traditional types of advertising and direct response marketing too, since the same principles around campaign profitability will apply, or to other similar forms of online advertising.
One variable which should be accounted for carefully is the Life Time Value and how this may change for different types of customer, who may have different buying preferences and patterns, hence resulting in higher or lower LTV’s. The usual way to address this is to apply customer segmentation – that is, breaking the whole customer base down into different groups or segments, with members of each segment sharing similar characteristics with one another, but being sufficiently different from the other groups to mark them out as separate.
Where sales and other activities are taking place online, segmentation has become increasingly easy to track and apply using cookies or by advertising to different segments differently, applying different media/offer codes, driving them to different landing pages, and so on.
By segmenting, we can analyse data and behaviours of these different groups of customers, and in turn modify how we interact with them, the kinds of offers we make to them, the products we market, etc. all with a view to maximising our sales both within each segment, and overall across the whole business.
For example, some customers only purchase our highest value products. Even when we offered a cheaper product they didn’t take it. They generate high recurring revenues and have high life time values, but we would lose money if we extended the marketing to them to offer product ranges they weren’t interested in. Conversely, some customers have only ever purchased our lower-priced offerings, and even then they tend to wait for the sales or other discounted offers to be made before they’re persuaded to purchase. Here there’s no purpose in wasting money promoting our big-ticket items as they won’t be likely to buy them. But across all the customers within each group or segmentation, we can optimise our marketing, promotions and pricing, so optimising our overall sales.
The flip side is those customers who we really are better off without, they cost us money and administrative time and cost in servicing yet they buy only one product (probably the cheapest one), and then complain or never stick around to buy anything from us again. We are probably spending as much to recruit this sub-par customer as we are recruiting our best customers. So we will also want to optimise the kinds of prospects we market to and seek to acquire.
Segmentation analysis focusses on identifying who are our best customers, and what kind of prospects they were when we first acquired them so we can identify and recruit more customers like them. In order to do this kind of analysis, we’ll need to collect a number of different descriptive segmentation metrics, such as:
- What is the source of each visitor to the website? (online ads; offline ads; organic search; link referrals; joint venture/affiliate referrals; social media referrals – either our own social media marketing activity or activity by others; email campaigns; links from press/PR campaigns; direct)
- What device did each visitor use to visit the website? (mobile – iOS vs. Android; desktop; operating system; browser)
- Geographical location of each visitor (country, state, city)
- New or returning visitor; registered or unregistered on our website
- Activity upon visiting the site (bounce; remain; duration of stay; number of page views)
- Tracking the path – clickstream – each visitor takes through the website
Significant differences in relative metrics can guide the company towards ways to optimise results, e.g. visitors coming from an organic search have a significantly lower bounce rate than those coming from paid ad’s, therefore maximise resource usage on improving and optimising for organic traffic (through search engine optimisation – SEO) and deprioritise pay-per-click advertising.
Some quick SEO basics:
Content should be current and up-to-date, substantive and relevant to the subject(s) being covered. Landing pages shouldn’t be diluted by talking about multiple unconnected subjects. Obtain back-links – incoming links from other sites to your site – from authority sites with a high reputation (such as mainstream media sites, highly-ranked blogs, high quality review sites). Increase your social media presence on platforms like Facebook, Twitter and Google+, post quality original content there and encourage likes, retweets and shares – especially shares and retweets by influential people with high reputations.
References:
[1] Business Metrics for Data-Driven Companies, Duke University