"...TalkTalk’s former CEO Dido Harding interviews Alexander Nix, CEO at Cambridge Analytica". — Transcript

Full transcript

1. 0:04i'm very excited about the fact that
2. 0:07rather than being bombarded with
3. 0:08messages on products and services
4. 0:10or even political messages that don't
5. 0:13necessarily resonate with you or aren't
6. 0:15relevant to you as an individual
7. 0:17you can start to receive advertising and
8. 0:20communications
9. 0:22that are more relevant to your lifestyle
10. 0:24your economic means
11. 0:26your ideology and so forth
12. 0:34so alexander thank you so much for
13. 0:37joining me
14. 0:37for this discussion about how the
15. 0:39digital world is
16. 0:40changing everything i wonder if we could
17. 0:43start off by
18. 0:44you telling us a bit about what
19. 0:46cambridge analytica does
20. 0:47certainly thank you dido um so we're a
21. 0:50data-driven
22. 0:51marketing and communications agency
23. 0:53which really means that we're
24. 0:55we're trying to use data and predictive
25. 0:57analytics to identify and target
26. 0:59audiences such that we can start to
27. 1:02serve more
28. 1:02personalized content to um to the people
29. 1:06we're trying to engage
30. 1:07so quick google search will flag up that
31. 1:10you've used those skills and techniques
32. 1:12a lot in the political arena around the
33. 1:14world
34. 1:15what is it that you're doing with data
35. 1:17analytics that's driving the
36. 1:19effectiveness of various political
37. 1:21campaigns
38. 1:22well it's actually a great shame that we
39. 1:25we consider ourselves a technology firm
40. 1:27that happens to apply its
41. 1:30technologies and methodologies in a
42. 1:32number of theaters including
43. 1:33politics and defense and government
44. 1:35space but principally in the brand and
45. 1:37commercial space but
46. 1:38certainly politics is what we've
47. 1:40received most press for
48. 1:42uh in the last uh six or twelve months
49. 1:44because of
50. 1:45the elections that we've been involved
51. 1:47with
52. 1:48to answer your question in the political
53. 1:50sphere
54. 1:51what we try and do is use data to
55. 1:54understand
56. 1:55what messages are most important to
57. 1:58individual
58. 1:59constituents to individual voters such
59. 2:01that we can align
60. 2:02the politicians policy and
61. 2:06ideological standpoint with the voters
62. 2:09so that they can be well informed about
63. 2:12what what position the candidates are
64. 2:14taking
65. 2:14and therefore have enough information to
66. 2:16make up uh their minds about who they
67. 2:18want to vote for
68. 2:19can you give us a sense of the scale of
69. 2:21personalization
70. 2:22say in a campaign that you've been
71. 2:24involved in how
72. 2:25what proportion of the population might
73. 2:27you be
74. 2:28directly specifically communicating with
75. 2:31as opposed to sending a generic message
76. 2:32to
77. 2:33well it actually depends very much on
78. 2:35the legislative framework of the
79. 2:37countries that you're working in so
80. 2:38different countries
81. 2:40uh are more or less permissive in their
82. 2:42data legislation
83. 2:43countries such as the uk is what's known
84. 2:46as an
85. 2:47opt-in data environment so people have
86. 2:50to give their consent for their data to
87. 2:51be used for these types of programs
88. 2:53america is an opt-out so people have to
89. 2:56obviously
90. 2:58contact you and say that they don't want
91. 3:00their data used so in a country such as
92. 3:03america were able to target all adults
93. 3:06some 230 million adults voters
94. 3:10or registered voters across the country
95. 3:12and to be able to start to engage them
96. 3:14with specific messages within the
97. 3:17campaign
98. 3:18so just for those of us that are not so
99. 3:20technically literate that means that for
100. 3:22every single citizen in the u.s
101. 3:24you've amassed a raft of data and
102. 3:27specifically targeted a unique message
103. 3:29to
104. 3:29to them well the idea is to try and look
105. 3:32at
106. 3:33what data there is available on all the
107. 3:36adults across a certain country and then
108. 3:39to
109. 3:39look at correlations into that in that
110. 3:41data such that you can
111. 3:43begin to cluster people people
112. 3:46who have similar attitudes or lifestyles
113. 3:48or
114. 3:49geographies or demographics to each
115. 3:52other
116. 3:52so that you can begin to serve different
117. 3:54clusters of people with different
118. 3:56messages that are going to be most
119. 3:57relevant to them
120. 3:59and why is it that it seems to be the
121. 4:02outsiders the non-mainstream political
122. 4:05candidates that have been so quick to
123. 4:07adopt these sorts of techniques
124. 4:09i'm not sure that's true actually i mean
125. 4:12if you look in america it was originally
126. 4:14the democrats who started to explore the
127. 4:16use of digital and then big data and
128. 4:18analytics
129. 4:19clearly we were we were working very
130. 4:21closely with the rnc
131. 4:22the republican national committee which
132. 4:24is obviously a main political body
133. 4:26um i think um
134. 4:29a lot of political parties are beginning
135. 4:31to invest in these technologies
136. 4:33certainly in more established political
137. 4:35parties you do see
138. 4:36legacy methodologies being embraced by
139. 4:41by people who maybe have been entrenched
140. 4:43in the system and
141. 4:44and are still fighting yesterday's
142. 4:46campaign and therefore maybe are a
143. 4:47little bit less
144. 4:48uh willing to explore new technologies
145. 4:50maybe in smaller fringe parties
146. 4:52uh they're a bit more dynamic um than
147. 4:55the larger ones that's probably the same
148. 4:57in business effectively
149. 4:58often it is the new entrance or the
150. 5:00challenges like the business that i ran
151. 5:02that is willing to try new things
152. 5:03exactly that okay so you see the same
153. 5:06phenomenon across in the uk as well or
154. 5:07not
155. 5:08i think that um i think in the uk
156. 5:12we're seeing um i mean certainly both in
157. 5:14the last general election both
158. 5:16labour and the conservatives and other
159. 5:18political parties started to invest
160. 5:20quite heavily and embrace these
161. 5:21technologies
162. 5:22so i don't think uh you could say that
163. 5:24it's been confined to to more fringe
164. 5:26parties uh
165. 5:26maybe such as ukip which received a lot
166. 5:28of press uh
167. 5:30for its use of data and digital yeah i
168. 5:32was thinking less maybe
169. 5:33my mischief was less fringe parties and
170. 5:36more non-mainstream candidates when
171. 5:38barack obama first
172. 5:39started campaigning and using probably
173. 5:41he's probably the first
174. 5:43global globally recognized politician to
175. 5:45be
176. 5:46using digital as a platform wasn't he he
177. 5:48was not a
178. 5:49you know mainstream expected winner in
179. 5:51the the
180. 5:53the democratic primaries initially was
181. 5:54he indeed but i think
182. 5:57you've got to look at um maybe not look
183. 6:00at the candidates you've really got to
184. 6:02look at what's happening outside of
185. 6:03politics
186. 6:04you know silicon valley was for the
187. 6:06first time uh beginning to
188. 6:08develop technologies uh and embrace
189. 6:10these technologies that really weren't
190. 6:12possible to implement
191. 6:13before about 2008 and certainly 2012.
192. 6:16you know the idea of using data in such
193. 6:18a granular way
194. 6:20simply wasn't possible we didn't have
195. 6:22that capability
196. 6:23that we do today so these are still
197. 6:25quite new there's going to be a first
198. 6:26first mover advantage clearly the united
199. 6:29states which is developing a lot of
200. 6:30these
201. 6:31tech it is going to have the edge there
202. 6:34but now we're seeing a lot of european
203. 6:36uh countries embracing these sorts of
204. 6:38technologies across the commercial space
205. 6:40and of course
206. 6:41entering the political uh space as well
207. 6:43so is there anything that you can do
208. 6:46that you shouldn't well i'd like to
209. 6:49think not
210. 6:50i mean obviously technology has um uh
211. 6:53you know has limitless bounds in some
212. 6:57sense but
213. 6:57we're all bound by some moral code and
214. 7:00ethical code and fair business practice
215. 7:02and and the same applies to politics so
216. 7:05um
217. 7:05you know anyone can can bend the laws uh
218. 7:08or break the rules
219. 7:09but um that isn't really a reflection of
220. 7:12technology that's a reflection of the
221. 7:13individuals
222. 7:14i was more thinking is there anything
223. 7:15that the technology is enabling possible
224. 7:18that as a society we shouldn't allow so
225. 7:21for example
226. 7:22in was it 19 i think it was 1957
227. 7:25subliminal advertising was banned
228. 7:28is there anything akin to that in the
229. 7:31data analytics techniques
230. 7:33that that you see developing that we
231. 7:36should be thinking about regulating
232. 7:38well i think at this particular juncture
233. 7:40i see it
234. 7:41as a very positive evolution i'm very
235. 7:44excited about the fact that
236. 7:45rather than being bombarded with
237. 7:47messages on products and services
238. 7:49or even political messages that don't
239. 7:52necessarily resonate with you or aren't
240. 7:53relevant
241. 7:54to you as an individual you can start to
242. 7:56receive um
243. 7:58advertising and communications uh that
244. 8:01uh
245. 8:01are more relevant to your lifestyle your
246. 8:03economic means um
247. 8:05your ideology and so forth so i think
248. 8:07that is positive it's positive for the
249. 8:08consumers or the voters
250. 8:10they get a better um more targeted
251. 8:12experience it's clearly better for the
252. 8:14for the brands or the political parties
253. 8:16because they get a return on investment
254. 8:18is improved
255. 8:20um but i think obviously you have to be
256. 8:23careful that it doesn't become intrusive
257. 8:25at the moment the data that's being used
258. 8:27is is fairly benign
259. 8:29um companies like ours are modeling
260. 8:32uh consumer and lifestyle and
261. 8:34transaction data you know
262. 8:36what cars you drive what magazines you
263. 8:38read this is this is
264. 8:40not necessarily the sort of data that
265. 8:42most people
266. 8:43would feel um
267. 8:46made them vulnerable or leave them left
268. 8:49and exposed
269. 8:50in some sort of way i think as long as
270. 8:53that line is maintained
271. 8:54and we're not talking about financial
272. 8:57data or sensitive
273. 8:58personal data or health data or anything
274. 9:00else that people might find intrusive
275. 9:02then most people would agree that
276. 9:05the benefits outweigh the risks