Preserving Personalized Advertisements for More Accurate Web Archives

Posted by Mat Kelly on July 31, 2024 · 4 mins read

COVID Web Ad >

Advertisements on the web are annoying to experience yet when preserved, are useful for analyzing our digital pasts. In collaboration with Old Dominion University’s Web Science and Digital Libraries (WS-DL) Research Group (@WebSciDL), we at Drexel CCI (@DrexelCCI) are pleased to announce that the follow-on to our “Saving Ads…” (see blog post) project, “Preserving Personalized Advertisements for More Accurate Web Archives”, has been selected for funding from IMLS starting in August 2024!

The project will run for 2 years and focus on preserving personalized advertisements on the web—the ones that are tailored to individuals or specific demographics—with the realization that preserving dynamic, much less personalized web content, is difficult and frequently not performed.

Schedule of Completion >

IMLS Logo >

In this project spanning from August 2024 to the end of July 2026, we will explore research questions like:

  1. To what extent do institutional web archives capture personalized advertisements on the web?
  2. Do scholarly and lay users prefer (re)using archived web pages that include personalized ads, that include a generic comprehensive capture, or that do not include web pages with ads?
  3. How might the strategic use of diverse personas illuminate web content that would otherwise go unarchived?

Our continued focus on diversity will further resonate with this project as we seek to preserve more accurate representations of past web pages than what you might have seen as captured from a web archive’s crawler. This project will consist of four phases including tasks for evaluation that will allow us to circle back to prior phases with further informed exploratory processes after the initial investigations. Each of the respective phases will focus on (1) Persona development and data collection, (2) Evaluation of data relative to web archives’ holdings, (3) Evaluation of captures based on users’ expectations, and (4) Results dissemination and technical web archive supplementation.

Location-based ad Personalized Ad
Two examples of personalized ads: one is location-based (a 2024 concert in Virginia for a CNN.com page from 2013) and one is based on browsing history (a current NVIDIA graphics card for a 2014 article).

SAA Logo > iPRES 2024 Logo >

Some initial findings from the preliminary planning grant of this project, so far, have been presented at the 2024 Society of American Archivists (SAA) Research Forum (slides) and will be presented at the upcoming 2024 International Conference on Digital Preservation (iPRES) this September. Further, additional technical details on our initial exploration are being finalized for dissemination as we bridge our IMLS Planning and upcoming Applied Research Grant.

Additional information about the project is available directly from IMLS with specific details, use cases, planning, progress, and other details to come as the project progresses. Thank you again, IMLS, for the continued support of the “Web Ads” project by Drexel CCI and ODU WS-DL!