Modeling Personalization Investment Costs and Overall ROI / Value

A look at investment costs associated with implementing personalization and calculating our overall ROI and value.

Modeling Personalization Investment Costs and Overall ROI / Value

This is part four of a four part blog series.

In my previous blog posts in this series, we looked at understanding the core value of our digital property, followed by modeling the benefits of personalization. Now let’s look at investment costs associated with implementing personalization and calculating our overall ROI and value.

personalization 1

First, estimating investment costs: We need to look at implementation project costs and ongoing operational costs associated with personalization. I typically use a 3 year timeframe to calculate Total Cost of Ownership (TCO). The cost categories for implementing personalization are roughly the same as those for building and maintaining a traditional website. The chart below illustrates splitting build and maintain against creative and tech.

personalization chart

Just with building a website, there are no hard and fast personalization figures for each cost category above. Some organizations may allocate costs 1 & 2 as one-off project costs (Capex), whereas others may smooth these costs, absorbing them into 'business as usual operational expenditure (Opex). But broadly speaking, these categories stand as a useful way to separate and estimate costs. I go into more detail on this below, in the example.

Once we’ve modeled the investment costs, we can simply move to Step 4 which involves subtracting the costs from the benefits, to arrive at an ROI. Again, this is typically measured over a 3 year timeframe, in order to calculate the total net benefits, or the project’s net economic return. Finally, we can detailed the non-financial impact to provide an overall value case, as outlined in my first blog in this series.

So, now time for an example.

Step 1: Let’s take a mid-size retailer with an annual revenue $300m / yr. Roughly 12% of revenue, or $36m, comes directly from the commerce website.

Let’s assume this is the baseline value pre-personalization; $36m per year, or $3m per month.

  • We can see the site has 2m visitors per month and 60,000 transactions per month. A conversion rate = 3%
  • The average online revenue per user (ARPU) is $50, with 60,000 transactions per month, giving $3m revenue.
  • Sales are seasonal so the exact number varies by month. What is interesting here is not necessarily the exact numbers but more the relative numbers before and after personalization. So, we need to ensure we’re controlling for any seasonal, or other, differences when it comes to testing our hypotheses, that personalization will increase conversion rates and revenue.
  • The site also advertises third party products and makes $4m / yr in revenue generation from ad revenue and referrals.
  • In addition, the site has a lead gen / awareness / brand building element and also has some product support information, although we’re not quantifying that value here.

Now, let’s assume the retailer has grown 5% each year, so we’ll fairly model a 5% growth each year regardless of whether we implement personalization. Ad revenue remains flat. We could model something like this:



Step 2: Now, let’s assume we implement personalization in the form of content targeting: displaying unique content to users based on something we know about them such as context, behavior or explicit identification. Let’s say, we aim to segment skiers and hikers and display content based on location, and sport preference. We might see the following results.

No impact to number of unique visitors per month or acquisition channels (sources - mobile traffic, organic traffic, click events, referring URLs). This shouldn’t be surprising as we haven’t done anything to drive new traffic to the site or change SEO.

However, we see slight improvements on engagement metrics for the content targeted sample:

  • Number of page views (pages per session) increases 5%
  • Time on site increases by 10%
  • Bounce rate decreases by 10%
  • Exit rate decreased by 10%

Overall, we see revenue increase in two ways; The engagement metrics correlate to an increase in conversion rate (from 3% to 3.5%) and the ARPU increases by 10% (from $50 to $55).

So, overall with personalization, conversion on 2m visitors is now 70,000 transactions per month at $55 = $3.85m rather than $3m in Yr 1. Sales are seasonal, so we need to ensure we’re measuring like for like, when looking at any uplift for personalization but we can also test against a control group which sees general, not personalized content.

Also, given the improved engagement metrics, we can show a 10% increase in ad revenue. So, when we forecast projected revenue with personalization against that without, we get the following comparison chart - showing forecast revenue before and after personalization.


The total difference in revenue over 3 years = $33,355,500.

Step 3: Investment cost to implement personalization

In this example, the typical costs for illustrative purposes, are outlined here:

  1. Creative - Personalization strategy & use cases. Executed through 1-2 workshops with the site owners and editorial team. A third party facilitated this, with one-off $15k consultancy cost, within a timeframe of 1-2 months. Internal costs up total costs to roughly $50k.
  2. Tech build / design & configuration of the personalization tool (driven by functional requirements). This cost category can vary wildly; It can include the tool purchase, together with the tech cost to setup and configure the tool. Here the tool is Cloud and subscription based, so relatively simple to configure. The initial tech investment required was around $15k of external consultancy to advise / configure + internal costs, totaling $30k.
  3. On-going personalization / including managing metrics / analytics. Typically a marketing team might have a site owner, content editorial, tech representatives, data, analytics.. Here, I’ll roughly model 1 FTE equivalent at $120k / yr fully loaded cost - and assume a 10% increase year on year.
  4. Technical maintaining of the tool with updates. Assuming a Cloud based subscription model for the tool itself, we can round to $100k annual subscription cost for the tool. A single tech FTE ($100k/yr loaded cost) may own this and other marketing tech. If we can assume 20% of their time is dedicated to managing personalization we arrive at an annual $20k effort cost + $100k subscription cost = $120k annual opex.

This gives us a table with a 3 year personalization TCO as follows:


It is worth noting that I have seen personalization initiatives run as small projects for significantly less than this. I’ve also seen truly transformative omni-channel personalization initiatives for orders of magnitude more than the sum shown here. This is for illustrative purposes only.

Step 4: In this example, is this investment worth it? This step is a case of simple maths, subtracting investment costs from benefits to get net benefits. When we chart the benefits minus costs, we get the following:


In other words, set-up will cost around $80k and TCO over 3 years = $837k.

This investment realizes a 3 year net revenue increase of $32.518m. So, the net revenue benefit is 38.8x the investment, a healthy ROI.

Now it is important to consider this is a revenue based model. Some organizations may want to model profit rather than revenue for their true ROI. Let’s assume in retail, the profit margin on revenue is 25%, then the total net profit benefit = $8.13m. Now the ROI on pure profit = 9.71x the investment.

Step 5: The ROIs here are impressive. A sound and complete business case always considers more than just the numbers and sometimes it is the soft benefits which will justify a personalization initiative, especially if the numbers alone are underwhelming, as can be the case for smaller organizations. Increases in customer retention, enrichment, and advocacy have not been modelled here but may make an initiative a no-brainer, outside of ROI.

Finally, it is sometimes interesting to note what other organizations are doing. I’ve read a number of articles recently stating that personalization in 2016 is a must-have, not nice to have. A number of organizations are implementing personalization - not based on the numbers but FOMO (fear of missing out). What’s the risk of doing nothing? Perhaps a simple break even on the numbers is enough to justify the investment, given the other intangible benefits.

If you found this blog interesting and would like to complete a value assessment for implementing personalization on your digital properties, please get in touch.

Featured Resources

View More Resources