Angelika Dobrieva

BAIT Biotechnology · Case study

The second site is the real test of the first

The client I redesigned Hemogenyx for came back with a second company and a one-line brief: build it like Hemogenyx.

Role
UX/UI designer — end to end
Client
BAIT Biotechnology — second commission from the same client
Status
Design delivered · site not yet public
My scope
System reuse · Page structure · New components · Responsive UI · Handoff

01 — From One Brand to Two

Reading “make it like Hemogenyx” correctly

A repeat commission is the most useful feedback a designer gets, and not for the obvious reason. It is not praise. It is a test with a pass mark.

The pass mark

  • If it was a systemThe second costs a fraction
  • If it was a layoutThe second costs the same again

Three readings of the brief

  • Layout copyInherits answers to problems it does not have
  • Template fillWhatever does not fit gets cut from the story
  • System instanceWhat transfers transfers, the rest is redesigned

Only the third reading survives contact with the content. So the interesting work here was not drawing screens. It was deciding, component by component, which half of the first site I was allowed to keep.


02 — From a Listed Company to a Pre-Launch One

The same visual language, genuinely different jobs

Two companies

  • Hemogenyx — listedFilings · Share data · Governance · Registrar
  • BAIT — pre-launchA leadership team

Half of the first site's navigation existed to serve a shareholder who already knew which PDF they wanted. BAIT has no such visitor — no filings, no registrar, no announcements.

Navigation

  • Nine-plus items
  • Four items

The site gets faster rather than emptier. And at pre-launch the people are the credibility, so the team moves from a strip on About to a destination of its own.


03 — From Reuse to an Audit

Sorting components by what they answer

The sorting rule

  • Answers a shape of contentCarries
  • Answers a particular clientRedesigned

Carried across untouched

  • Carried
    • Floating pill navigation — repeated in the footer
    • One oversized word opening every page
    • Two-column reading blocks
    • Oversized figures with a small coloured label
    • A circular outline button closing every page
  • One component, two jobs

    → a visitor learns the pill shape once and meets it again in the footer

  • Orientation before content

    → on a site where most pages look alike at a glance, the oversized word says where you are

  • The number carries the argument, the label carries the caveat

    → a statistic block that works for any company's evidence

  • A page never ends on nothing

    → there is always one named next step

Redesigned

  • Laboratory photographySoft gradient motif
  • Investor architectureFour navigation items
  • People as a stripPeople as a page

A pre-launch company has no image library, and stock photography would have undone the credibility the typography is building.


04 — From a Sub-Page to a Destination

The one component the first site never needed

BAIT's leadership spans four groups a visitor might want separately, on one page. Hemogenyx had no equivalent, so this had to be designed rather than reused.

The taxonomy

  • People
    • R&D team
    • Board of advisors
    • Board of directors
    • Alumni

The interaction

  • All people
  • Filter
  • Grid
  • Overlay biography
  • Filter pills across the top, with All People as the resting state

    → the taxonomy is visible before it is used; nobody has to guess the groups exist

  • Role label above the photograph, name below it

    → the grid scans by function rather than by face, which is how an investor reads a team

  • Cards on a fixed column width

    → “Director Pre-Clinical Development” wraps without pushing the row out of alignment

  • A biography opening in an overlay rather than a new page

    → the visitor keeps their place in the grid instead of navigating back to it

The trade-off

  • Keeps the visitor's place in the grid
  • Has no address of its own
Overlay backed by a real URL

For a pre-launch company whose leadership is its strongest asset, a biography that cannot be linked to is a real cost — an investor who wants to forward one CV has nothing to forward. I would raise it before build rather than after.

Still unanswered

  • Who decides when somebody moves to Alumni
  • What a filter shows when its group is empty
  • Whether counts belong on the pills

Those are maintenance questions, and a taxonomy with no rule for them drifts within a year — the same way an unmaintained component library does.


05 — From a Duplicated File to a Safe Launch

The deliverable the original never needed

The risk

  • Duplicate the file
  • Fast
  • The previous company's words stay

Working from an existing site is fast, and the speed is exactly what makes it dangerous. Everything survives until it is explicitly replaced.

Where they hide

  • Inherited content
    • Body copy
    • Footer registration line
    • Page titles and meta descriptions
    • Image alt text
    • The team
    • Therapeutic claims from the other pipeline

The deliverable

  • Content-replacement checklist
  • Signed off
  • Public

Tied to handoff rather than to design. That is the part of this project I would defend hardest in a review — not because it is visible, but because it is the failure mode nobody budgets for.

Status

  • Design delivered
  • Site not yet public
  • No analytics in scope

Reflection

Building the same thing twice tells you which decisions were principles.

Designing one site, it is impossible to tell your principles from your preferences. Everything you did feels necessary, because everything you did was in response to something. The second brand separates them in an afternoon: the decisions that answered a shape of content lift out cleanly and keep working, and the ones that were really about that particular client stop making sense the moment the name above them changes.

The pill navigation carried. The investor architecture did not, and should not have. Knowing which was which before starting is what made this project cheap rather than merely fast.

The uncomfortable half of the lesson is in section 05: a duplicated file is a liability until the last inherited word is gone. Speed came from reuse, and the only honest way to take that speed is to pay for it with a checklist.

Next case
Definder