Modeling ROI for Personalization: Valuing Your Digital Properties
This is part two of a four part series
In my previous post, I introduced the concept of crafting an ROI for personalizing your digital properties. Now it’s time to dive into more detail.
The first step is to understand your core business and digital customer experience as it relates to personalization. This is potentially the hardest step, so bear with me.
To measure success of implementing personalization, we need to be clear on what we’re measuring, how we’re measuring it and to ensure any changes we see are as a result of our actions. This step in the process is required as a baseline from which to demonstrate a financial uplift after implementing personalization.
First ask; what is your website worth? This doesn’t mean how much would it cost to build, but rather, how much revenue is it generating? The answer to this question should be a prerequisite when considering any enhancements to your digital properties. That said, answering this question can be far from straightforward, especially for non-commerce sites.
For digital properties that aren’t generating revenue directly, how do you articulate the operating value? Wikipedia lists 50 different types of websites ranging from; awareness sites, blogs, to brand-building sites, galleries and wikis, in addition to commerce. How do we measure the dollar value of all these?
The first critical step is to go back to the core business. Ask, how do you make money, then ask, how does the digital property support this core business? The following are the four most commonly encountered business models of a digital property:
- Sites selling something (direct revenue), e.g. commerce, subscription sites
- Sites aiming to sell something indirectly, e.g. content for commerce, lead generation sites
- Awareness sites including news or publishing. These mostly rely on advertising revenue and occasionally subscription revenue
- Sites for savings, e.g. process efficiency, help, support or FAQ sites.
Often different business models merge across a single site or group of sites. Sites can also satisfy both B2C or B2B and often also vary by geography, business unit, region etc. This complicates things and is partly why the task of quantifying value can prove to be difficult. The Venn diagram below illustrates the potential complexity of this situation.
The most common business models spanning typical websites. Where does your site fit?
Given this potential complexity, my recommendation is to start by focusing on a single site business model - that is, just one non-overlap area of the Venn diagram. Obviously getting hard quantifiable value is easier for some business models than others. The site goals will vary by type of site and most product owners measure site metrics according to these goals. If you’re wondering what metrics should matter the most, Kate Fogarty, Sr. Manager, Digital Marketing at Acquia, has put together a quick guide on which web metrics a digital marketer should care about and why. These metrics all have their place but for measuring ROI we want a hard link between site metrics and quantifiable value - i.e. money.
I’ve provided some examples here of metrics to measure, depending on your site type.
Commerce & Subscription Sites
The value of commerce and subscriptions sites -- sites selling something -- is linked directly to revenue. The link is so strong one can measure the value per minute, hour, or day. For example:
- When Amazon went down in 2013 for a reported 40 minutes, the outage was estimated to cost $4.72 million in lost sales, based on the company’s average sales of $117,882 per minute.
- When the British supermarket - Sainsbury's - site went down in 2008 due to an internal glitch, the site was valued between £569k and £710k per day. This calculation was based on 8,000 to 10,000 shoppers a day, with an average basket size (average revenue per user, or ARPU) of £71.
Typically, a commerce site owner will report on additional, detailed metrics to provide general indicators of business health. These can also provide a baseline for measuring uplift from personalization and include:
- For projected revenue look at the total number of customers, conversion rates and lifetime value (LTV) of a customer. Calculating LTV can vary - here’s a good short blog for further information on the value of new customers
- Average Revenue per User (ARPU), also referred to as Revenue per Visitor (RPV), or Average Order Value (AOV)
- Monthly Recurring Revenue (MRR) and Annual Recurring Revenue (ARR) for subscription sites
- Cost to acquire a Customer (CaC) and Churn.
In short, measuring value from commerce and subscription sites is a simple link to revenue and / or we can measure additional metrics impacting revenue and profit.
Lead Generation & Content for Commerce
Content marketing sites create valuable content as a means to build awareness, but ultimately as a motive to increase customers and revenue. These sites are aimed at pushing people through the sales funnel but don’t necessarily include a commerce engine, or capture the final stages of the funnel. Despite the lack of a direct commerce component, these sites still play a significant role in the sales process. Modeling value, however, can be a little more challenging. This is where I have a lot of conversations around baseline value of a site prior to implementing personalization.
Let’s take an organization with a general lead generation site. A prospect may be aware of the brand via general marketing, TV advertising and / or word-of-mouth referrals. The prospect may receive a mailshot, an email, and go online to the website and eventually buy in-store. How is the website’s influence quantified across the varying channels?
We are now in ambiguous territory. What role does the site play in awareness, education, driving conversions, preventing sales drop outs etc? Whilst the websites isn’t generating income directly it is influencing sales and therefore has value.
For retailers, both online and traditional, Forbes claims that 30% of shoppers begin their searches on Google while 49% start on Amazon. The vast majority of shoppers in total begin a search online, when you include sites such as Etsy, eBay and Shopify. So, what effect does the site have, or not have, in this commerce journey even when the purchases are primarily off-line?
If we can measure how leads convert into customers and we can calculate lifetime value (LTV) of a customer, we can estimate revenue and value: Total leads x conversion x LTV
But this assumes we’re accurately measuring leads and conversion rates. Sites help build brand awareness but the link from brand awareness to ‘leads’, or ‘conversions’ and therefore revenue isn’t always clear.
Sure, we can look at engagement metrics such as; number of visitors, unique visitors, page views, time on site, bounce rates, exit rates etc. But these don’t necessarily relate to revenue either.
One option is to baseline site goals, such as the percentage of visitors who clicked on a page (CTR), or filled in a form, or submitted a query, or even a customer satisfaction survey. We can then look at uplift in these quantifiable metrics after implementing personalization and make some assumptions on value from here.
As an example, we can hypothesize that 30% of buyers visit the site at some point in their decision journey, even if the final purchase is in store. We can also assess the customer decision journey as see ‘drop-outs’ at various stages. We may hypothesize that 25% site visitors would go elsewhere if they couldn’t find the information they were looking for online. So, we can start to attribute a given percentage of revenue to the site, albeit assumptions based. We may attribute a range of say, 2% to 5% or whatever % of revenue to the site, as long as the underlying assumptions are reasonable and can perhaps be tested.
Crafting ROI is often a bit of both art and science and this is definitely the more arty part.
Publishing sites are often driven primarily by advertising revenue and are usually measured on engagement metrics, such as: unique visitors, page views, time on site, ‘impressions’ (i.e. CPM) or clicks (i.e. CPC) or some combination thereof.
Showing improvements in engagement scores will often drive value in advertising revenue. This can often be quantified relatively easily. When Google’s home page was offline for about five minutes August 2013, estimates suggested this cost Google about $545,000 in ad revenue.
Interestingly, some awareness sites are moving towards alternative revenue streams, based on monetising user data. Think Facebook and LinkedIn.
- Facebook, a market cap of $253 billion and the ninth-biggest company in the S&P 500. Clearly Facebook has a value based on user data which is far higher than its advertising revenue potential.
- LinkedIn recently sold to Microsoft for $26.2 billion. The value is in the data rather than revenue generated from recruitment consultants and users paying for services.
The key is to understand the model and how much the data is ultimately worth. If you know the value of your customers, or the value of your customer data, you can work out a monetary value attributable to users of your website.
Some websites can digitize processes in such a way to unlock significant value by compressing timelines and eliminating duplication or inefficiencies. A simple example many finance companies opt for includes converting customers to receive on-line documents rather than hard paper copies. This can reduce the cost of printing and posting, resulting in significant savings for both the business and customers; a win-win.
Many simple information-only / support websites can save costs. Common questions can be answered in on-line FAQs rather than from a call center. Reducing the number of enquiries made to a call center using an online support website can save significant costs. This is especially true in government and public sector services.
Example metrics we’d look to measure as a baseline here include:
- Support calls / on-line tickets resolved
- Engagement metrics; Time on site and CTR. Interestingly, these metrics are often reversed to show reduced, rather than increased, time on site and shorter click-through rates to show the site is channeling the user to the appropriate resolution quicker
- Customer satisfaction ratings.
The True Value of Your Site
Any of the models above often coexist on the same site and to complicate things, some sites include alternative revenue streams, or new innovative models which don’t easily fit into the above categories. The key is to carefully consider the different ways that your website contributes to your business and aim to quantify that. It is only realistic to track improvements in something when that ‘something’ is quantified and regularly monitored.
As a final check point, you can cross-check the operating value of your site against the spend on your site. This is a way to ‘mark your own homework’ when determining site value.
Gartner suggests the average spend on marketing is around 10% of total enterprise revenue. If we assume 25% of the 10% (or 2.5%) of total revenue is spent on digital / online / maintenance of the corporate website, then this is the value you’re assigning to your site. If the value you’re calculating is less than the spend allocated above, you may not be calculating true value, or you’re overspending on your site!
Another view is to think about the negative impact on your business if your website shutdown tomorrow. What would the cost be to your business? This is the true value of your site.
In my next blog I will show how we can build on this baseline to help justify the investment of implementing personalization on your site.