Having worked in search marketing for a few years, I was curious to see if there was an easy way to measure the latent impact (delayed conversion) of search marketing on online sales/consumer actions.
As marketers we sometimes forget that the prospect converts at his/her own convenience and not ours. Sometimes, the prospect may convert right away and sometimes the prospect may convert at a later date.
I was unable to find a rule of thumb on how many conversions may happen on a delayed basis. This is why I was delighted when I had the opportunity to run a campaign on Yahoo for a change instead of Google-only campaigns.
Latent Impact Data Gathering (Google versus Yahoo):
While Google adds back the “delayed” sale into the sales/conversions reported for the day that a user first clicked on to a search ad, I think Yahoo does a smarter job in this respect. Yahoo reports the sale/conversion on the actual day the “delayed” sale occurs. This may well be 3, 4, 5 or even 30 days after the user clicked on a search ad.
In the case of Google, one has to re-run the reports to see how sales/conversions for the previous days/weeks contributed to lowering the earlier recorded Cost per Order. In addition to that, one has to download and store the historical data offline, and then compute the latent impact across several days.
Results:
The Yahoo campaign produced some interesting results in comparison to Google.
Overall, the cost per order/conversion was much lower on Google than on Yahoo when the campaigns were live.
Here are some top line stats with respect to the test program that I ran on Yahoo:
Program dates – April 7 to May 1, 2009
- The first orders on Yahoo came in on April 13 – however, in the case of Google, the first conversions flowed in on day 1.
- Of all the conversions reported – 69% of the conversions happened during the time frame the campaign was “live”.
- A respectable 31% of the conversions occurred after the campaign was over. This clearly demonstrates that search marketing has a measurable latent impact on the mind of the prospect. The last sale/conversion came in on June 7, a good 37 days after the campaign was considered over.
- The Cost Per Order reported after these additional 37 days helped decrease the CPA by 31%. Once I factored this “delayed/latent” Cost/Order, Yahoo as a search marketing channel was not far off from Google as in terms of cost efficiency.
Conclusion & Recommendation:
Search Marketing has become a hostage to its own success. Clients expect a strong positive ROI from the get-go and are unwilling to wait for “all results” to come in. As online marketers too, we must keep coming up with better and more complete ways of measurement.
Based on my experience, in this case, one of the easiest things we should do is to wait before we report the “final” numbers on a campaign.
I am hoping that by sharing this information, fellow online marketing professionals will have a third party stat to convince clients or in house marketers to budget for at least 1 in 3 responses/conversions coming in after the campaign is over and record success accordingly. If anyone has some other similar stats, please share it with us below.
Photo credit: #15: Buy something online by garethjmsaunders
Hi Harmit,
Nice article. I agree with you regarding the importance of measuring latent conversions. However, it is hard to come up with a rule of thumb.
In our case we always run reports (for more than 100 campaigns) back for a 30 day period to measure the impact of conversions.
Depending on the industry and seasonality, we see various percentages. Latent conversions seem to be more prominent in B2B industry and in the case of high ticket consumer products.
During Christmas season, people seem to be more impulsive as we see less latent conversions.