Tech

Project 6: Software phase 2.1 and 2.2: closing report

Project 6, aka Software Phase 2, was written up in three parts:

  • Phase 2, the original pitch
  • Phase 2.1, which was a significant change to the original Project 6 Software Phase 2; and
  • Phase 2.2.

In Project 6 (Software Phase 2), we originally set out to add several new features we believed would be crucial or valuable for paperight.com:

  • automatic PDF layout (to cope with the increasing load of incoming content)
  • mobile-optimised catalogue/browsing site with white-labelling feature
  • integration of major retailer affiliate-link feeds
  • publisher dashboard for sales figures and ability to edit own content
  • incorporate tracking URLs that enable single-sign-on forum access.

Between that pitch and briefing the developers, we changed our priorities based on user feedback and strategic considerations. This resulted in a significant change, where we revised our priority features as follows, referring to them as Software Phase 2.1:

  • 3rd Party Product Integration
  • Product Ownership
  • Publisher Dashboard: Transactions, Payments, Credits widget
  • Find-outlets call to action
  • General Improvements
    • Show PDF file size
    • Country to currency mapping
    • Average default print costs
    • Licensee details capture
    • SEO friendly URLs
    • Suggest a book
    • Licence search
  • PDF Automation
    • HTML to Paperight PDF
    • PDF to Paperight PDF

In 6B/Phase 2.2, we pitched to develop additional, closely related features, referred to collectively as Software Phase 2.2:

  • Site-content management (admins can change text on the site)
  • Preview-document downloads
  • Company User Management: separate ‘company’ registrations form ‘user’ registrations
  • Paperight Staff User Management (different user privileges)
  • Ability to Close/Reopen Accounts
  • Removal of advert space in PDF
  • Hide product-ordering and pricing details from non-logged-in users
  • Revised home page layout
  • HTML to Paperight PDF: further dev on epub imports

General report-back

Overall, this project went very well, thanks to good preparation on our side and great professionalism by Realm Digital. We did change priorities for development during the project, but this was understandable since we need to respond to user feedback as it comes in, especially if it affects our development roadmap while we’re developing.

As with other projects, our only concerns are that we and Realm should find ways to do more testing on older versions of IE (7 and 8) under different circumstances — but that can only really be achieved by seeing the site working on the field and responding.

All of the features developed in this project work well, and almost all have already proven useful or critically important in practice.

Objectives achieved

Planned: “These are important improvements to our product and our infrastructure. In particular: they reduce confusion and uncertainty for users (preview docs, find-outlets); they enable us to refine our messaging continuously in response to user feedback and context (e.g. emphasise topical content, or test different calls to action); they speed up prep of and improve the quality of our documents (epub-to-Paperight PDF); they make clear the distinction between individual users and the companies they represent — companies, not individuals, register on Paperight and deal in licences, and we must make this clear for user clarity and for legal clarity. We want these improvements to lead to higher conversion rates (more copy shops completing transactions, more customers going into Paperight outlets) as a result of clearer messaging, confidence from previews, and more content availability (as a result of our faster content-prep workflow).”

These objectives were all achieved and have proved very important to our work and strategy.

Objectives not achieved

None.

Measures of success

Original pitch

Planned Outcome
Are all our features implemented on time and do they work as planned? Yes.
Do they enable us to learn the lessons we want to learn, particularly around revenue streams? In part. We moved white-label site functionality to a later development phase (phase 3), and have not seen revenue increase there because Postnet are still delaying head-office-level signup. And we’ve now tested and proven that affiliate income (Amazon links) is not a revenue source. So while we have no positive outcomes from the experiments, the lessons have been valuable.
Expect: Ten-fold increase in the rate of adding new content (currently about 50 books a week); Reports from outlets of increased consumer-initiated interest; Some revenue from affiliate links within two months of that feature being live; Publishers updating their own metadata (e.g. descriptions and pricing). As explained further below, increase in rate of adding new content has been closer to two-fold.

We have seen greater consumer-initiated interest at outlets.

No revenue from affiliate links.

We decided not to allow publishers to update their own metadata, and have not seen demand for that feature anyway.
Like: In addition, an increase in rate of outlets and small publishers signing up abroad Outlet signups abroad have not been significant (two or three outlets abroad). However, we’ve seen several small and medium publishers sign up abroad, which has been great.
Love: One large South African publisher contribute a range of books (or a single major text prescribed at a major college or university) because of the availability of single-sign-on tracking URLs. This has not happened yet.

 

Project 6B pitch (revised original pitch)

 

Planned Outcome
Are all our features implemented on time and do they work as planned? As with Phase 1 there was some delay over original planned schedule. All features work as planned.
Expect: Ten-fold increase in the rate of adding new content (currently about 50 books a week); Previews being downloaded, and leading to some full purchases; Increased rate of purchased material We have not seen a ten-fold increase. The increase is likely a doubling of speed, which is still very useful. This is largely because it has proved near impossible to automate dealing with the technical differences between content sources, and some human intervention is still required.

Previews are being downloaded; it is not yet clear whether they directly lead to more sales, but anecdotal evidence suggests they do prevent sales of the wrong documents, reducing our customer service burden.

We see an overall increase in purchase rates that may be in part attributed to these software improvements, but can’t be separated from improvements in other areas of the business.
Like: In addition, an increase in rate of outlets and small publishers signing up abroad; Some revenue from affiliate links within two months of that feature being live We’ve seen one or two overseas signups, but no significant increase.

We have seen no revenue from affiliate links. This was not a good use of our resources, though it seemed a very good idea at the time. Worthwhile experiment, though, since we can now exclude affiliate revenue as a revenue source.
Love: One large South African publisher contribute a range of books (or a single major text prescribed at a major college or university) as epubs, and these books being purchased in outlets. The large publishers that have signed up have only given us PDFs so far, because they do not have EPUBs of the books they are giving us. This may still happen in future.

 

Budget

Original budget: R592160.00

Actual spend: R553 641.00

Returned to pool: R38519.00

Item Budget Actual Return to pool Comments
Software phase 2.2 572609.00 534090.00 38519.00 Software expenses up until October 2014 to be carried over to a new budget
Legal expenses 19551.00 19551.00 0
TOTAL 592160.00 553641.00 38519.00

 

Outputs and deliverables

 

Planned Outcome
Phase 2.1
3rd Party Product Integration Done. Search results show affiliate links to Amazon products. Has proven that this is NOT a source of revenue, and not as helpful as we hoped.
Product Ownership Done. Critical piece of the puzzle, allows us to link sales to publishers’ earnings automatically, in real time.
Publisher Dashboard: Transactions, Payments, Credits widget Done.
Find-outlets call to action Done.
Show PDF file size Done.
Country to currency mapping Done.
Average default print costs Done.
Licensee details capture Done.
SEO friendly URLs Done.
Suggest a book Done.
Licence search Done.
PDF Automation: HTML to Paperight PDF Done.
PDF Automation: PDF to Paperight PDF Done — extremely useful, and means we no longer require Adobe products in our workflow (except for rare and tricky exceptions)
Phase 2.2
Site-content management (admins can change text on the site) Done. Extremely powerful, has allowed us to fine tune wording from overall messaging to tweaking phrases on tooltips and buttons to increase site usability.
Preview-document downloads Done. Very useful and important feature.
Company User Management: separate ‘company’ registrations from ‘user’ registrations Done.
Paperight Staff User Management (different user privileges) Done.
Ability to Close/Reopen Accounts Done.
Removal of advert space in PDF Done. (Docs can now either support ad space or not.)
Hide product-ordering and pricing details from non-logged-in users Done.
Revised home page layout Done — crucial change, this, reduced dud registrations on the site from dozens a month to almost zero.
HTML to Paperight PDF: further dev on epub imports Done, but epub imports still present problematic issues, related to the wide range of epub structures, so we do not use this feature as much as we thought we would, preferring to work with the HTML extracted from the epub instead where possible.

 

Learnings

We and Realm should find ways to do more testing on older versions of IE (7 and 8) under different circumstances — but that can only really be achieved by seeing the site working on the field and responding. But Realm give us two weeks after going live to do final testing, after which most bug fixing is charged for over and above the project fee. The problem is that some problems (esp those with IE) only arise in the field months after release. This is an issue we’re discussing with Realm to find good compromises.

Exit/Sustainability/Viability

This was a once-off development phase that builds towards Paperight’s overall self-sustainability as a business.

Conclusion

This project went well overall. Some of the features we developed were experiments, and we learned valuable lessons from them even if they weren’t the lessons we were hoping for. We’re pleased with the changes as made midway due to user feedback received at the time.

Next steps

At the time of writing: for software, a phase 3 has been completed and phase 4 is starting.