Yes You Can Predict the Future of BPO | Article

Everest’s Predictive Model

fortune teller“In good times outsourcing is a good business; in bad times outsourcing is a great business.” Many of us have heard this phrase. At an intuitive level it makes sense, but has anyone really looked at the data to see if it is true? We decided to go ahead and do just that.

Growing at a hurried pace of an average of 38 percent per year for the last 10 years, BPO is currently in the “rapid-growth” stage on the much-talked-about “S-curve” of BPO adoption. The S-curve is expected to take decades to play out, given the sheer magnitude of the market, the relatively low current penetration, and the addition of new processes. Therefore, although we don’t think that the 38 percent growth can be extrapolated forever, BPO will remain in this phase for some time to come.

The actual number of BPO deals signed each year will not fit the “S-curve” as perfectly as a draftsman would draw the theoretical curve (see Exhibit 1). Therefore the question is: How do we explain deviations from the expected trend line?

Exhibit 1 — Expected BPO Adoption

graph 1

Our Hypothesis and Test

We used the old adage, “in good times outsourcing is a good business, in bad times outsourcing is a great business” as the basis of our hypothesis. We reasoned that variations from the BPO trend line should be explained at least in part by variations in the economy.

In order to provide a sufficient number of data points for the analysis, we looked at 413 global multi-process BPO deals1 including HR, procurement, and finance and accounting signed over the last ten years. For economic activity, we looked at US GDP as a proxy for economic activity affecting the buyers of outsourcing globally.

Both data series have a strong underlying trend line. Comparing any two data series with strong positive trends will always produce a correlation; therefore, we looked at variations from the mean for both GDP and BPO to avoid spurious correlations. We decided to look for an inverse relationship, reasoning that below-trend economic growth would lead to above-trend BPO growth. We also lagged the data by one year, reasoning that bad economic times cause buyers to begin looking at BPO, which result in deal signings the following year. For both GDP and BPO we looked at the most recent 10 years of data.

We didn’t cheat; that is, we didn’t do any data mining. Even before looking at the data, we established our hypothesis: that there would be an inverse relationship between GDP and BPO growth with a one-year lag.

Exhibit 2 — Year-to-Year Variations in GDP and BPO Growth Rates

We can see an inverse relationship between GDP and BPO growth with a one-year lag.

graph 2

Looking at the data (see Exhibit 2), we were immediately encouraged. Look at the GDP data. The low point in GDP growth occurs in 2001, and the high point in BPO growth occurs in 2002; that is, there is an inverse relationship with a one-year lag, just as our hypothesis predicted.

After seeing the encouraging pattern in the data, we ran the actual regression statistical analysis and were pleasantly surprised to find an R-squared of 70 percent; that is, statistically, 70 percent of the variation in the number of BPO deals from the BPO average growth rate can be explained by variations in the growth rate in GDP. For readers who are not statisticians, this is a strong correlation for a single variable regression model. See Exhibit 3 for a graphical look at just how good the fit is.

Looking ahead, based upon above-trend 2006 GDP growth, the model predicts a below-trend 32 percent growth in the number of global BPO deals for 2007.

Exhibit 3 — Predicted versus Actual BPO Deals

graph 3

Much additional analysis is needed. We have not yet answered questions such as:

  • How does the model perform for individual processes such as HR versus F&A?
  • How does the model work for individual countries?
  • Does the model work better for some countries than others?
  • Can the model be used to compare the expected BPO market for one country versus another?

For now, we are going to make the sweeping assumption that the model works not only in the aggregate but also by process and by country.

What Implications Does This Have for Buyers?

Potential buyers should consider the expected supplier response they receive when buying BPO services. A buyer approaching the market when the economy is in the doldrums will not get as much attention from busy suppliers as that same buyer will get during a year when the economy is strong. If you know that BPO is in your future, buy from a position of strength; don’t wait for an economic downturn like the rest of the herd.

What Implications Does This Have for Suppliers?

Sophisticated suppliers will adjust their marketing and sales efforts according to economic conditions. Our findings suggest that for any given year, you can measure the GDP growth for any country against the GDP average growth for that country and predict whether you will see an above- or below-average year for BPO signings in that country. Business development efforts/budgets should be allocated to the country with the greatest negative deviation from trend line GDP.

Furthermore, within a country, sales and marketing efforts should be directed to the economic sectors performing furthest below their long run sales and earnings trend lines.

1 Multi-process BPO deals include FAO, HRO, and PO contracts with minimum of two processes outsourced, over US$1 million in Annualized Contract Value (ACV), and a minimum contract term of 3 years

Lessons from the Outsourcing Journal:

  • The growth in BPO is impacted inversely by the growth in GDP.
  • Buyers, if they know BPO is in their future, should consider entering the market when the economy is strong to avoid the crowd.
  • BPO suppliers should focus their business development efforts in the markets with the worst economies.

Post a Comment

Your email address will not be published.

( required )

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>